The DOI matches its Crossref record; 212 of the 212 references it cites were found in a registry or an index; 63 of the 161 sources its claims cite were checked for support.
Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges
Open accessAbstract
A literature review of human-written, LLM-generated, attributed, and co-authored text detection problems, methods, and challenges.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This survey maps authorship attribution in the LLM era into four problem types and reviews methods, datasets, metrics, and open challenges. It argues that LLMs blur human–machine authorship boundaries and that progress requires better generalization, explainability, standardized benchmarks, and interdisciplinary work.
Key findings
- The paper organizes the field into four representative problems: human-written text attribution, LLM-generated text detection, LLM-generated text attribution, and human-LLM co-authored text attribution, with complexity increasing across them.
- Detection approaches are grouped into metric-based and model-based families, including feature-based, neural network-based, zero-shot, and watermarking methods; neural detectors often outperform metric-based ones but are less explainable, need retraining for new LLMs, and are vulnerable to adversarial or poisoning attacks.
- The survey compares 21 benchmarks/datasets: all support LLM-generated text detection, but fewer support LLM-generated text attribution or human-LLM co-authored attribution, and many treat human-written text as one class, making them unsuitable for identifying individual human authors.
- Human judgment is weak: untrained reviewers distinguish GPT-3 text from human text at roughly chance level, and experienced faculty/researchers identify GPT-generated academic writing correctly only about half the time. Commercial detectors often claim over 99% accuracy, yet few are tested on shared benchmarks and many have high false-positive rates.
Methods
- Design: a literature review/survey that synthesizes existing research rather than presenting new experiments.
- Data/sample: cited prior studies, 21 authorship attribution datasets/benchmarks, and a table of commercial and open-source LLM-text detectors.
- Analysis: qualitative thematic categorization of four attribution problems, comparison of methodologies, datasets, evaluation metrics, and discussion of future directions.
Limitations
- Stated: fewer studies address human-LLM co-authored attribution because it is difficult, and existing work often simplifies the task; many datasets are unsuitable for individual human authorship attribution; detectors lack robustness to decoding strategies, prompts, adversarial attacks, unseen domains, and unseen models.
- Stated: LLM-based attribution becomes less effective as the number of candidate authors grows due to context-length limits, and training-data biases can produce inaccurate attributions for marginalized groups and languages.
- Evident: as a survey, it provides no new empirical validation or standardized benchmark results, and its comparisons draw on heterogeneous prior datasets and metrics, so reported performance is not directly comparable; the text also does not describe a systematic search or inclusion protocol.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
27 Partly supported · 24 No full text of the sources · 18 Supported · 18 Own interpretation · 17 Sources not found · 10 Not checked yet · 4 No basis given · 3 Supported so far · 3 Couldn’t confirm · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“In contrast, detection algorithms frequently outperform the human in this task [Ippolito et al., 2020].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Supportedcites Ippolito et al., 2020 source
The Human Evaluation section reports that human raters, including trained experts, had consistently worse accuracy than automatic discriminators across all tested decoding methods and excerpt lengths. This supports the citing sentence’s comparison for the text-generation settings studied. (LLM, conf=0.94)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“untrained human reviewers are often unable to distinguish GPT-3-generated text from humanwritten text, identifying it correctly only at a rate consistent with random chance [Clark et al., 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Supportedcites Clark et al., 2021 source
Section 2.4 (Results; page not provided) reports that evaluators identified GPT-3 versus human authorship correctly 49.9% of the time, compared with 50% random chance, and that their accuracy was not significantly different from chance. The Conclusion identifies these evaluators as untrained. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The rapid advancements of Large Language Models (LLMs) have blurred the lines between human and machine authorship, posing significant challenges for traditional methods.”
- Also stated in: § 1. INTRODUCTION · plain
“The advent of LLMs has complicated authorship attribution, making it increasingly difficult to distinguish between LLM-generated texts and human-written texts [Clark et al., 2021;Sadasivan et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > Abstract
What the sources say
- Supportedcites Clark et al., 2021 source
The Results (§2.4) report that evaluators distinguished GPT-3 text from human-authored text at 49.9% accuracy, approximately chance, compared with 57.9% for GPT-2. The Analysis (§2.5) links that increased difficulty to improvements in newer models and raises concerns about existing human-evaluation methods. (LLM, conf=0.88)
- Supportedcites Sadasivan et al., 2023 source
The Introduction (p. 2) says LLMs produce human-like text and that this creates an authenticity challenge. The Abstract (p. 1) describes reliable AI-text detection as a challenge and reports vulnerabilities in existing detection methods, supporting the claim that distinguishing machine from human authorship has become difficult. (LLM, conf=0.90)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 1. INTRODUCTION · plain
Every source the paper cites for this claim was checked and supports it.
In the paper
“Existing detectors also lack robustness to various factors, such as alternative decoding strategies [Ippolito et al., 2020], input sequence length [Solaiman et al., 2019], different prompts [Kumarage et al., 2023;Lu et al., 2023], and repetition penalties [Fishchuk and Braun, 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.3 Open Challenges
What the sources say
- Supportedcites Solaiman et al., 2019 source
In “Our Work,” the paper says shorter outputs are harder to detect than longer ones. Figure 2 in “Roberta-Large Transferred Model Accuracy” shows detection accuracy increasing with text length, supporting the claim that detector performance is sensitive to input sequence length. (LLM, conf=0.91)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Sadasivan et al. [2023] provided theoretical insights indicating that the detection problem is becoming increasingly difficult.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Supportedcites Sadasivan et al., 2023 source
Section 4 (p. 12) explicitly says that, as more advanced language models more closely mimic human text, the detection problem becomes increasingly difficult under the paper’s theoretical result. The Introduction (p. 4) makes clear that this conclusion depends on the distance between human and AI text distributions decreasing. (LLM, conf=0.94)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“However, existing detectors and models for LLM-generated text are not yet fully reliable [Sadasivan et al., 2023;Wang et al., 2023;Dugan et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Supportedcites Sadasivan et al., 2023 source
The Abstract (p. 1) reports that paraphrasing attacks substantially reduce detection rates, exposing vulnerabilities in current AI-text detection systems. The Conclusion (p. 37) also warns that false positives can make their practical applications unreliable. These findings support the claim that existing detectors are not fully reliable. (LLM, conf=0.92)
- Supportedcites Wang et al., 2023 source
The Feasibility of Black-Box Machine-Generated Text Detection section says current approaches may be less effective and generalize poorly to unseen domains, generators, and languages. The paper’s empirical findings also describe detectors misclassifying machine-generated text in those settings. Page numbers are not supplied in our copy. (LLM, conf=0.92)
- Not checked — no full text of the sourcecites Dugan et al., 2024 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Humanauthored texts are generally more coherent, while LLMgenerated texts tend to repeat terms within a paragraph [Dugan et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.1 Featured-based Method
What the sources say
- Partly supportedcites Dugan et al., 2023 sourceExperimental — accuracy not yet measured
Section 6, “RoFT-Specific Insights” (page not available in our copy), reports that irrelevance and contradiction were among the most reliable reasons participants gave for identifying generated text. That lends some support to a coherence difference, but the paper does not establish that human-authored text is generally more coherent across LLMs, or that generated text tends to repeat terms within a paragraph; it records repetition as a participant-selected reason, not a measured paragraph-level tendency. (LLM, conf=0.78)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“These detectors also face limitations such as overfitting to training data [Uchendu et al., 2020].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.2 Neural Network-Based Detectors
What the sources say
- Could not verifycites Uchendu et al., 2020 sourceExperimental — accuracy not yet measured
The supplied passages contain bibliographic metadata, not a discussion of detector limitations. Searches for overfitting, generalization, and training-data limitations returned only the title and author listing (no section or page), so the available copy does not establish this claim. (LLM, conf=0.72)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“These detectors require retraining when encountering text generated by new LLMs to ensure reliable detection [Mitchell et al., 2023;Chakraborty et al., 2023b].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.2 Neural Network-Based Detectors
What the sources say
- Partly supportedcites Mitchell et al., 2023 sourceExperimental — accuracy not yet measured
Section 5.1 reports that supervised detectors can overfit to their training data, including the source model, and perform worse under distribution shift. Section 5 contrasts this with zero-shot detectors, which do not require retraining for new source models. The paper does not establish that neural network-based detectors invariably require retraining for every new LLM to ensure reliable detection. (LLM, conf=0.84)
- Not matchedcites Chakraborty et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Existing detectors also lack robustness to various factors, such as alternative decoding strategies [Ippolito et al., 2020], input sequence length [Solaiman et al., 2019], different prompts [Kumarage et al., 2023;Lu et al., 2023], and repetition penalties [Fishchuk and Braun, 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.3 Open Challenges
What the sources say
- Supportedcites Ippolito et al., 2020 source
The Transferability section reports that detectors trained on one decoding strategy often perform poorly when tested on another, including a top-k-trained detector whose accuracy fell to 42.5%. The Conclusion likewise says that discriminators generally transfer poorly between decoding strategies, while noting that training on a mix of strategies can help. (LLM, conf=0.93)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
- It may state two separate claims as one. 38
- Probably not a substantial claim. 42
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Additionally, detectors are vulnerable to adversarial attacks, including homoglyph attacks [Gagiano et al., 2021;Macko et al., 2024], whitespace insertion [Cai and Cui, 2023], syntactic perturbations [Bhat and Parthasarathy, 2020a], synonym replacement [Kulkarni et al., 2023], and paraphrasing [Krishna et al., 2024b;Shi et al., 2024;Becker et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.3 Open Challenges
What the sources say
- Not matchedcites Gagiano et al., 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Macko et al., 2024 source
Macko et al. (2024) is cited for the homoglyph-attack portion of the sentence. The Text edits section describes its homoglyph attacks, and Section 4.1 reports that these attacks often evade machine-generated-text detection across the tested languages. The paper also notes that preprocessing can mitigate them. (LLM, conf=0.94)
- Supportedcites Cai and Cui, 2023 source
The Introduction reports that adding a single space before a comma substantially reduces AI-text detection rates. Section 3, “Space Infiltration,” describes this whitespace-insertion attack as a way to bypass detectors. The other attack types in the citing sentence are attributed to other sources, not this paper. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
Every source the paper cites for this claim was checked and supports it.
In the paper
“Techniques like slanted triangular learning rates and gradual unfreezing can be used to avoid catastrophic forgetting during fine-tuning [Howard and Ruder, 2018].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.3 Pre-trained Language Models
What the sources say
- Supportedcites Howard and Ruder, 2018 source
The Introduction explicitly identifies slanted triangular learning rates and gradual unfreezing as techniques to retain previous knowledge and avoid catastrophic forgetting during fine-tuning. The Conclusion also says the proposed techniques, used together, prevent catastrophic forgetting. (LLM, conf=0.95)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Typically, this task is approached as a binary classification problem [Zellers et al., 2019;Solaiman et al., 2019;Jawahar et al., 2020b;Fagni et al., 2021;Mitchell et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.1 Problem Definition
What the sources say
- Supportedcites Zellers et al., 2019 source
Section 5 (p. 6) describes detecting machine-generated news by classifying articles as Human or Machine. Section 5.1 (p. 6) explicitly requires the discriminator to assign one of those two labels to each article in its unpaired evaluation, supporting the binary-classification characterization. (LLM, conf=0.88)
- Supportedcites Solaiman et al., 2019 source
The “Automated ML-based detection” section describes classifiers that distinguish language-model outputs from a true-text distribution. “Our Work” describes a detector trained on samples equally likely to be real or fake, supporting the characterization of detection as a binary classification task. (LLM, conf=0.89)
- Supportedcites Jawahar et al., 2020b source
The Introduction describes the commonly used approach as a classification task distinguishing machine-generated from human-written text. Those are two classes, matching the claim that detection is typically approached as binary classification. No page number is provided in the supplied passage. (LLM, conf=0.93)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Rivera-Soto et al. [2021] concluded that topic diversity and dataset size are crucial for effective cross-domain transfer.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.3 Pre-trained Language Models
What the sources say
- Partly supportedcites Rivera-Soto et al., 2021 sourceExperimental — accuracy not yet measured
The Conclusion attributes Reddit’s strong cross-domain transfer to both topic diversity and dataset size. But §4.2 says size only partially explains Reddit’s performance and that increasing the number of Amazon authors yielded marginal gains. Calling both factors “crucial” for cross-domain transfer generally is stronger than the paper’s finding. (LLM, conf=0.82)
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Authorship attribution methods are typically divided into classification-based methods for a small set of candidate authors and similarity-based ranking methods for larger numbers of authors [Rivera-Soto et al., 2021;Huertas-Tato et al., 2022].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.1 Problem Definition
What the sources say
- Partly supportedcites Rivera-Soto et al., 2021 sourceExperimental — accuracy not yet measured
The Introduction (page unavailable) contrasts a prior closed-world study of 21 authors with this paper’s open-world setting involving far more authors. The Experiments metrics passage (page unavailable) describes ranking candidates by cosine similarity. The paper does not say that authorship attribution methods are typically divided into classification-based and similarity-based categories, or that classification is generally used for small candidate sets. (LLM, conf=0.84)
- Partly supportedcites Huertas-Tato et al., 2022 sourceExperimental — accuracy not yet measured
The paper describes classifier-based authorship approaches in §2.1 (p. 3) and evaluates its own similarity-based attribution from a pool of authors in §5.2 (p. 8). It does not say that methods are typically divided into these two categories or associate classification with small candidate sets and similarity-based ranking with larger ones. (LLM, conf=0.78)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Beyond feature extraction, LLMs possess the ability to identify the author of unknown text based on nuanced linguistic features, making it possible to conduct end-to-end authorship attribution [Huang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
What the sources say
- Supportedcites Huang et al., 2024 source
The Introduction reports that LLMs can perform authorship attribution and use subtle stylistic and linguistic patterns to distinguish authors. The Conclusion describes applying end-to-end methods directly to authorship attribution while bypassing intermediate feature extraction. (LLM, conf=0.93)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Authorship attribution is a complex reasoning task, and LLMs possess significant capabilities in reasoning and problem-solving, particularly in zero-shot learning within resource-limited domains [Kojima et al., 2022].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
What the sources say
- Partly supportedcites Kojima et al., 2022 sourceExperimental — accuracy not yet measured
The Abstract (p. 1) reports substantial gains on diverse reasoning tasks using zero-shot prompting, and the Discussion (p. 9) describes LLMs as decent zero-shot reasoners with potential for generic problem solving. The paper does not establish these capabilities specifically within resource-limited domains; its zero-shot method avoids handcrafted examples and time-consuming fine-tuning, which is a narrower point. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Contrastive learning [Khosla et al., 2020] is often used with pre-trained language models to enhance stylistic representation by maximizing similarity between texts written by the same authors and minimizing it with texts from different authors [Huertas-Tato et al., 2022].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.3 Pre-trained Language Models
What the sources say
- Not checked — no full text of the sourcecites Khosla et al., 2020 source
- Partly supportedcites Huertas-Tato et al., 2022 sourceExperimental — accuracy not yet measured
Section 2.3 (p. 5) says the authors pair pre-trained transformers with contrastive learning to build representations of writing style. The Introduction (p. 2) says texts by the same author should have higher similarity than texts by different authors. The paper supports this approach, but its discussion of recent efforts does not establish that it is often used for stylistic representation. (LLM, conf=0.82)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“One variation of this task is to detect the boundary between human-written and LLM-generated text [Cutler et al., 2021;Wang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.1 Problem Definition
What the sources say
- Not matchedcites Cutler et al., 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Wang et al., 2024 source
Section 3.3, “Boundary Identification,” defines a task in which a human begins a text, an LLM continues it, and a detector identifies the boundary. The Introduction also lists human-written-to-machine-generated change-point detection as Task 3. This supports the cited description of a variation involving boundary detection. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Human-authored texts tend to be more coherent and exhibit greater lexical diversity compared to LLM-generated texts [Guo et al., 2023;Dugan et al., 2023;Zhang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.2 Methodologies
What the sources say
- Not checked — no full text of the sourcecites Guo et al., 2023 source
- Partly supportedcites Dugan et al., 2023 sourceExperimental — accuracy not yet measured
Section 6, “RoFT-Specific Insights” (page unavailable), reports that irrelevant and contradictory sentences were among the most reliable signs participants used to detect generated text. This lends support to the coherence comparison, but the paper does not establish that human-authored text has greater lexical diversity; its sentence-level analysis examines length, parts of speech, and named entities instead. (LLM, conf=0.83)
- Not checked — no full text of the sourcecites Zhang et al., 2024 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The evaluation of the quality of machine-generated excerpts has traditionally relied on human judgement, which is considered the gold standard for open-domain generation systems [van der Lee et al., 2019;Gehrmann et al., 2019a].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Supportedcites van der Lee et al., 2019 source
Section 2, “Automatic versus human evaluation,” explicitly says that human evaluation remains the gold standard for assessing overall NLG system quality. The Introduction describes human evaluation as crucial to assessing system quality, supporting the claim’s broader account of its role. The paper makes this point about NLG generally rather than singling out open-domain systems. (LLM, conf=0.91)
- Could not verifycites Gehrmann et al., 2019a sourceExperimental — accuracy not yet measured
The Abstract reports a human-subjects study of detecting generated text, and the Method describes statistical tests for that detection task. Neither passage says that human judgment has traditionally been the gold standard for evaluating the quality of open-domain generation; searches for human evaluation and quality found no passage making that point. (LLM, conf=0.73)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Some studies categorize any text that is generated, modified, or extended by a machine as LLM-generated.This simplifies the task to either LLM-generated Text Detection or LLMgenerated Text Attribution [Yang et al., 2023c;Crothers et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.1 Problem Definition
What the sources say
- Not matchedcites Yang et al., 2023c
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Partly supportedcites Crothers et al., 2023 sourceExperimental — accuracy not yet measured
The Definition and Scope section defines machine-generated text as text produced, modified, or extended by a machine. Section IV describes detection as distinguishing machine-generated from human text, and Section V-B discusses attribution to a generative model. The paper uses the broader category “machine-generated text,” however; it does not categorize all such text as LLM-generated. (LLM, conf=0.86)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Some studies have formulated a seven-class classification that includes one human class and six LLM classes [He et al., 2023;Wang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.2 Methodologies
What the sources say
- Not matchedcites He et al., 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Wang et al., 2024 source
Section 5.3, “Multi-way Detection,” explicitly describes Task 2 as seven-label classification and lists one human class and six LLM classes. This matches the classification attributed to Wang et al. (2024). (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Models like DNA-GPT [Yang et al., 2023a] and DetectGPT [Mitchell et al., 2023] utilized the T5 model [Raffel et al., 2020] to simulate scenarios in which humans modify LLMgenerated texts.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.2 Methodologies
What the sources say
- Not matchedcites Yang et al., 2023a
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Mitchell et al., 2023 source
Section 5.2 states that DetectGPT's authors simulated human revision of machine-generated text by replacing word spans with T5-3B samples. This supports the claim as it pertains to DetectGPT; the separate citation to DNA-GPT is not assessed here. (LLM, conf=0.95)
- Not matchedcites Raffel et al., 2020
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
Every source the paper cites for this claim was checked and supports it.
In the paper
“In human authorship attribution, where there are a large number of candidate authors, retrieval metrics such as Mean Reciprocal Rank (MRR) and recall-at-k are used [Rivera-Soto et al., 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.2 Evaluation Metrics
What the sources say
- Supportedcites Rivera-Soto et al., 2021 source
The Experiments metrics passage, following §3.3, explicitly says the paper assesses performance with recall-at-8 and mean reciprocal rank (MRR). The Introduction describes its setting as open-world, with several orders of magnitude more authors than a 21-author study, supporting the large-candidate context. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The increasingly large context length of LLMs enables in-context learning (ICL) [Brown et al., 2020] by incorporating more documents, enhancing the model's ability to capture intricate writing patterns.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.5 Leveraging LLM Capabilities
What the sources say
- Partly supportedcites Brown et al., 2020 sourceExperimental — accuracy not yet measured
The Introduction (p. 5) describes in-context learning with as many demonstrations as fit in the model’s context window, and §3.7 (p. 20) reports improved performance with more examples in context. These passages support the role of context capacity in ICL, but they do not establish that incorporating more documents enhances the model’s ability to capture intricate writing patterns. (LLM, conf=0.85)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The challenge remains to provide higher-level explainability that aligns with human cognitive processes [Rudin, 2019].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.3 Explainability
What the sources say
- Supportedcites Rudin, 2019 source
In ‘Challenge #3’ (page unavailable), Rudin frames interpretability around how people explain their reasoning in visual classification and calls for models that reason about image parts so their explanations reflect the decision process. The Introduction also relates interpretable model design to limits on human cognitive capacity. This supports the claim’s broad point, though Rudin advocates inherently interpretable models rather than post-hoc explanations of black boxes. (LLM, conf=0.85)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Advanced techniques are required to dissect these integrations, identifying where and how the contributions of LLM intersect with human input [Wang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.3 Open Challenges
What the sources say
- Supportedcites Wang et al., 2024 source
Section 3.3, “Boundary Identification,” describes detecting where human-written text changes to LLM-generated text. The Limitations section says real mixed texts can be more complicated and calls for methods that detect mixed text and identify all changing points, supporting the claim’s need for more capable techniques. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Models like BERT and RoBERTa have shown limitations in cross-domain tasks [Barlas and Stamatatos, 2020b;Huertas-Tato et al., 2022], and adapting models to new domains remains difficult due to factors like dataset size variability and writing styles.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.2 Generalization
What the sources say
- Not matchedcites Barlas and Stamatatos, 2020b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Partly supportedcites Huertas-Tato et al., 2022 sourceExperimental — accuracy not yet measured
Section 2.1 (p. 3) says adaptation to new domains is usually hard and identifies variation in document size and style; Section 5.2 (p. 8) shows the RoBERTa baseline underperforming other models on out-of-domain authorship tasks. The paper does not establish a cross-domain limitation for BERT specifically, and its discussion of varying size concerns documents rather than explicitly identifying variability in dataset size as a cause of adaptation difficulty. (LLM, conf=0.84)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The Linguistic Inquiry and Word Count (LIWC) tool exemplifies how automated text analysis can use more than 100 psychological dimensions to analyze word use, reflecting distinct language variations among different groups in specific genres and languages [Goldstein-Stewart et al., 2009;Pennebaker et al., 2015;Dudău and Sava, 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.7 Integrating Interdisciplinary Perspectives
What the sources say
- Not checked — no full text of the sourcecites Goldstein-Stewart et al., 2009 source
- Not matchedcites Pennebaker et al., 2015
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Partly supportedcites Dudău and Sava, 2021 sourceExperimental — accuracy not yet measured
The Introduction describes LIWC2015 as an automated word-counting tool with categories that include psychological constructs, and the Discussion reports differences among the languages studied. But the Introduction says it has approximately 90 categories in total—not more than 100 psychological dimensions—and the cited passages do not establish the claim about variations among groups in specific genres. (LLM, conf=0.88)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“The challenge in social media stems from the brief nature of posts and a large pool of potential authors, making the attribution of short documents particularly difficult [Aborisade and Anwar, 2018;Seroussi et al., 2014;Theophilo et al., 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
What the sources say
- Not checked — no full text of the sourcecites Aborisade and Anwar, 2018 source
- Supportedcites Seroussi et al., 2014 source
The Abstract states that attributing online texts such as blogs and e-mails is more challenging than attributing traditional texts because they tend to be short and have more candidate authors. The Related Work section likewise identifies many candidate authors and short e-mails as challenges. (LLM, conf=0.94)
- Not checked — no full text of the sourcecites Theophilo et al., 2021 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“Misattribution can lead to wrongful accusations or misinterpretation of an author's intent [Lund et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 8. ETHICAL AND PRIVACY CONCERNS
What the sources say
- Could not verifycites Lund et al., 2023 sourceExperimental — accuracy not yet measured
The 'Authorship attribution is a major issue' section discusses ownership, plagiarism, and the difficulty of tracing sources in AI-generated text (page unavailable). Searches for wrongful accusations and misinterpretation of an author's intent found no passage addressing either consequence. (LLM, conf=0.72)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Could not verify
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“For instance, Kowalczyk et al. [2022] detected GPT-2-generated fake reviews using Shapley Additive Explanations (SHAP) [Lundberg and Lee, 2017].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.3 Explainability
What the sources say
- Not checked — no full text of the sourcecites Kowalczyk et al., 2022 source
- Supportedcites Lundberg and Lee, 2017 source
The Abstract (p. 1) identifies SHAP as “SHapley Additive exPlanations” and describes it as a framework for interpreting model predictions. The citation to Lundberg and Lee identifies the method named in the sentence; the fake-review study is attributed separately to Kowalczyk et al. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Although many of these detectors claim over 99% accuracy in detecting LLM-generated text, few are tested on shared benchmark datasets.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
- Rests on an assumption: Despite their high accuracy claims, many detectors suffer from high false positive rates, which could falsely accuse individuals of plagiarism and undermine credibility and trust in genuine authors.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“These detectors are primarily designed for LLM-generated text detection (Problem 2).”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“However, fewer support LLM-generated text attribution (Problem 3) and Human-LLM Co-authored Text Attribution (Problem 4).”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Neural network-based detectors generally outperform metric-based methods in both human authorship attribution and LLM-generated text detection problems [He et al., 2023;Zhang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 1. INTRODUCTION
What the sources say
- Not matchedcites He et al., 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not checked — no full text of the sourcecites Zhang et al., 2024 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
In the paper
“Chakraborty et al. [2023a] employed theoretical analysis to argue that detecting LLM-generated text is nearly always feasible with the collection of multiple samples, and they established precise sample complexity bounds for this detection.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Not matchedcites Chakraborty et al., 2023a
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
In the paper
“LLM-generated texts are typically less emotional and more objective than human-written texts [Guo et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.1 Featured-based Method
What the sources say
- Not checked — no full text of the sourcecites Guo et al., 2023 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“whereas deep learning methods demand significant computational resources and extensive labeled data, with the risk of catastrophic forgetting [Ramasesh et al., 2022].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.3 Open Challenges
What the sources say
- Not matchedcites Ramasesh et al., 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, these neural network approaches often offer less explainability compared to their metric-based counterparts.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 1. INTRODUCTION
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“The rapid evolution of LLMs further complicates detection, as newer versions exhibit different stylistic characteristics, making detection models quickly obsolete [Chakraborty et al., 2023b].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.3 Open Challenges
What the sources say
- Not matchedcites Chakraborty et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
In the paper
“Barlas and Stamatatos [2020a] found that BERT performed well with large vocabularies, outperforming multi-headed RNNs.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.3 Pre-trained Language Models
What the sources say
- Not matchedcites Barlas and Stamatatos, 2020a
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
In the paper
“Unlike BERT-based models, which require computationally expensive fine-tuning and large amounts of domain-specific data for optimal performance, LLMs can generalize across various domains without fine-tuning, thereby mitigating issues related to domain specificity [Barlas and Stamatatos, 2020a].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
What the sources say
- Not matchedcites Barlas and Stamatatos, 2020a
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
In the paper
“Collaborative texts, which blend multiple writing styles, further mask individual contributions and obscure distinct authorial signals [Dauber et al., 2017].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.3 Open Challenges
What the sources say
- Not checked — no full text of the sourcecites Dauber et al., 2017 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Traditionally, LLMs have been employed mainly for auxiliary tasks such as feature extraction and data annotation [Patel et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
What the sources say
- Not checked — no full text of the sourcecites Patel et al., 2023 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“These methods do not require hand-crafted features but require substantial training time and domain-specific labeled data, struggling with cross-domain generalization and explainability.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.3 Pre-trained Language Models
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The natural language understanding capability allows LLMs to recognize nuances, styles, and patterns in language, which are crucial for distinguishing between authors.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
In the paper
“LLMs are also effective with shorter texts, reducing the necessity for long inputs to derive meaningful representations [Eder, 2015].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.2.4 LLM-based Methods
What the sources say
- Not checked — no full text of the sourcecites Eder, 2015 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“Zero-shot detection methods are generally statistics-based, enabling the detection of LLM-generated text without additional training [Su et al., 2023b].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.3 Zero-Shot Detectors
What the sources say
- Not matchedcites Su et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This attribution task extends beyond binary classification to handle multiple classes, increasing the complexity of LLMgenerated text detection.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.1 Problem Definition
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“In contrast, LLMgenerated texts typically exhibit consistent syntax and a broader vocabulary [Guo et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.3 Open Challenges
What the sources say
- Not checked — no full text of the sourcecites Guo et al., 2023 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, LLMs may inadvertently reproduce snippets of their training data.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.2 Methodologies
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Feature-based methods are more explainable but have drawbacks, such as poor generalizability of certain features across different domains and sampling methods.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2.1 Featured-based Method
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“For instance, a 10-class classification might include seven human classes and three LLM classes [Tripto et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.2 Methodologies
What the sources say
- Not checked — no full text of the sourcecites Tripto et al., 2023 source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“These texts may begin as human drafts that are later extended or revised by LLMs, or vice versa, resulting in an integration of styles that obscures authorial markers [Liu et al., 2023b].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.3 Open Challenges
What the sources say
- Not matchedcites Liu et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Attributing texts to LLMs versus human writers involves recognizing inherent differences in their text generation capabilities.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.2 Methodologies
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim.
None of the sources cited for this claim have been checked yet.
In the paper
“As a result, the detection of LLM-generated text has garnered significant attention [Kumarage and Liu, 2023;Tang et al., 2023a;Wu et al., 2023b;Yang et al., 2023c].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3. LLM-GENERATED TEXT DETECTION
What the sources say
- Not checked — no full text of the sourcecites Kumarage and Liu, 2023 source
- Not matchedcites Tang et al., 2023a
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Wu et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Transformer-based models, such as BERT and RoBERTa, are fine-tuned on datasets containing both human-written and LLM-generated texts to conduct the attribution.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.2 Methodologies
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, LLMs can be misused to generate content that conceals true authorship by mimicking others or using LLMs to alter their work.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.3 Open Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This variability complicates authorship attribution, necessitating the identification of consistent and unique stylistic markers.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 2.3 Open Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may state two separate claims as one.
In the paper
“Variations in training data, model architecture, and fine-tuning processes contribute to the distinctive outputs of different LLMs, though these differences are often subtle and difficult to detect [Uchendu et al., 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4.3 Open Challenges
What the sources say
- Not checked — no full text of the sourcecites Uchendu et al., 2021 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“LLM-generated text detectors can be categorized into metricbased and model-based methods [He et al., 2023], which are further divided into feature-based, neural network-based, zero-shot-based, and watermark-based methods.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Not matchedcites He et al., 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Watermarking-based methods typically fall under the white-box detection category, while proprietary models are restricted to black-box methods.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“MIXSET [Zhang et al., 2024] offers a more comprehensive dataset that includes text refined by LLMs through polishing, completion, and rewriting operations.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.2 Methodologies
What the sources say
- Not checked — no full text of the sourcecites Zhang et al., 2024 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“LLM-generated text attribution builds upon techniques for LLM-generated text detection.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 4. LLM-GENERATED TEXT ATTRIBUTION
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“These detectors are also classified as white-box or black-box, depending on their access to the LLM weights [Tang et al., 2023b;Yang et al., 2023c].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 3.2 Methodologies
What the sources say
- Not matchedcites Tang et al., 2023b
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Yang et al., 2023c
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
In the paper
“Moreover, leveraging LLMs for data augmentation and synthetic data generation can create diverse training datasets, which in turn improves the generalization of attribution models [Albalak et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.5 Leveraging LLM Capabilities
What the sources say
- Not checked — no full text of the sourcecites Albalak et al., 2024 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A claim is checked only against the cited paper’s full text, never its abstract, and none is available for its sources.
In the paper
“Factors influencing the performance of existing authorship attribution models include the size of the training text [Hirst and Feiguina, 2007;Marton et al., 2005], the number of candidate authors [Koppel et al., 2006], and the imbalanced distribution of training texts among candidate authors [Stamatatos, 2008].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
What the sources say
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“Some detectors like GPTZero [Tian and Cui, 2023] can also detect human-LLM co-authored text and identify portions likely to be LLM-generated, addressing the boundary detection sub-problem of Problem 4 [Cutler et al., 2021].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
What the sources say
- Not matchedcites Tian and Cui, 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Cutler et al., 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This involves detecting subtle shifts in style and contextual cues that indicate the extent of LLM involvement, allowing for accurate segmentation and attribution of authorship within hybrid documents.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.3 Open Challenges
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“Authorship attribution is inherently multidisciplinary, encompassing elements of linguistics, computer science, forensic science, and psychology [Stamatatos, 2009].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.7 Integrating Interdisciplinary Perspectives
What the sources say
- Not checked — no full text of the sourcecites Stamatatos, 2009 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Additionally, Self-BLEU are useful metrics, with a lower score indicating higher textual diversity [Zhang et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.2 Evaluation Metrics
What the sources say
- Not checked — no full text of the sourcecites Zhang et al., 2024 source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“For example, tweets are limited to 280 characters, whereas legal judgment documents contain thousands of words [Seroussi et al., 2011].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
What the sources say
- Not matchedcites Seroussi et al., 2011
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Since Problems 2, 3, and 4 frequently treat human-written text as a single category for simplicity, rather than identifying individual authors as in Problem 1, many of these datasets are unsuitable for Problem 1, which requires unique author identification.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Combining these interdisciplinary perspectives enhances our ability to distinguish between human-and machine-generated texts, addressing the emerging challenges posed by sophisticated language models.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 7.7 Integrating Interdisciplinary Perspectives
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: fails — It may be too vague to check against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. It may be too vague to check against evidence.
In the paper
“Additionally, these detectors often lack robustness against variations in sampling strategies, adversarial attacks, and unseen domains and language models [Dugan et al., 2024].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
What the sources say
- Not checked — no full text of the sourcecites Dugan et al., 2024 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper relies on this without support.
The proprietary nature of many LLMs restricts access to comparative data, posing significant hurdles.
“However, the proprietary nature of many LLMs restricts access to comparative data, posing significant hurdles.”
The paper relies on this without support.
The high degree of stylistic overlap among LLMs, especially those with similar architectures or trained on overlapping datasets, further complicates accurate classification.
“Additionally, the high degree of stylistic overlap among LLMs, especially those with similar architectures or trained on overlapping datasets, further complicates accurate classification.”
The paper relies on this without support.
Hybrid texts may originate as human-written content, with LLMs employed for conditional generation, making it difficult to clearly distinguish between the stylistic features of human and machine contributions.
“Hybrid texts may originate as human-written content, with LLMs employed for conditional generation, making it difficult to clearly distinguish between the stylistic features of human and machine contributions.”
The paper relies on this without support.
There are fewer studies on human-LLM co-authored text attribution due to its difficulty, and existing research often makes simplifications.
“There are fewer studies on this task due to its difficulty, and existing research often makes simplifications.”
The paper relies on this without support.
The complexity of human-LLM co-authored text attribution increases as users may employ multiple LLMs to compose different sections of an article, further blurring the lines between human and machine-generated content.
“The complexity of this task increases as users may employ multiple LLMs to compose different sections of an article, further blurring the lines between human and machine-generated content.”
The paper relies on this without support.
Analyzing and classifying texts co-authored by humans and LLMs presents a significant challenge due to the blending of human and machine stylistic features.
“Analyzing and classifying texts co-authored by humans and LLMs presents a significant challenge due to the blending of human and machine stylistic features.”
The paper relies on this without support.
Human-LLM co-authored texts often start as human drafts and are later extended or revised by LLMs, or the process might occur in reverse.
“These texts often start as human drafts and are later extended or revised by LLMs, or the process might occur in reverse.”
The paper relies on this without support.
The integration of styles in human-LLM co-authored texts creates a hybrid form that makes it difficult to distinguish distinct authorial markers, thereby complicating the attribution process.
“This integration of styles creates a hybrid form that makes it difficult to distinguish distinct authorial markers, thereby complicating the attribution process.”
The paper relies on this without support.
Human-LLM co-authored texts pose a more intricate challenge due to the blending of stylistic and linguistic elements from both human authors and LLMs.
“Human-LLM co-authored texts pose a more intricate challenge due to the blending of stylistic and linguistic elements from both human authors and LLMs.”
The paper relies on this without support.
Manually collecting authorship data is time-consuming and costly, motivating researchers to utilize web data sources such as Wikipedia and Reddit.
“Manually collecting data is time-consuming and costly, motivating researchers to utilize web data sources such as Wikipedia and Reddit.”
The paper relies on this without support.
Custom datasets for LLM-generated text are easier and more affordable to create, and are often built alongside human-written text to maintain a similar domain and format.
“In contrast, custom datasets for LLM-generated text are easier and more affordable to create, and are often built alongside human-written text to maintain a similar domain and format.”
The paper relies on this without support.
LLM-generated and human-written texts are often disproportionate, making class imbalance a crucial issue.
“Addressing class imbalance is crucial, as LLM-generated and human-written texts are often disproportionate.”
The paper relies on this without support.
The rapid growth of online communication has led to shifts in writing behavior, resulting in shorter, fragmented, and less coherent social media tweets and text messages.
“However, the rapid growth of online communication has also led to shifts in writing behavior, resulting in shorter, fragmented, and less coherent social media tweets and text messages.”
The paper relies on this without support.
Despite their high accuracy claims, many detectors suffer from high false positive rates, which could falsely accuse individuals of plagiarism and undermine credibility and trust in genuine authors.
“Despite their high accuracy claims, many detectors suffer from high false positive rates, which could falsely accuse individuals of plagiarism and undermine credibility and trust in genuine authors.”
The paper relies on this without support.
Current authorship attribution methods face limitations when handling a more extensive range of candidate human authors or LLMs.
“Current authorship attribution methods face limitations when handling a more extensive range of candidate human authors or LLMs, presenting opportunities for future research.”
The paper relies on this without support.
Existing approaches for LLM-generated Text Attribution typically manage only a limited number of authors or models, which restricts their applicability in real-world scenarios where the pool of potential authors or models can be vast.
“For instance, existing approaches for LLM-generated Text Attribution typically manage only a limited number of authors or models, which restricts their applicability in real-world scenarios where the pool of potential authors or models can be vast.”
The paper relies on this without support.
Domain generalization poses significant challenges due to variability in vocabularies, syntax, and styles across different subjects, complicating accurate authorship attribution.
“Domain generalization poses significant challenges due to variability in vocabularies, syntax, and styles across different subjects, complicating accurate authorship attribution.”
The paper relies on this without support.
Authors' adaptation to various genres dilutes their identifiable stylistic traits, complicating attribution.
“Authors' adaptation to various genres dilutes their identifiable stylistic traits, complicating attribution.”
The paper relies on this without support.
Traditional attribution methods struggle to distinguish between human-authored texts and those generated by LLMs, which adeptly replicate these features.
“Traditional attribution methods, which rely on stylistic and linguistic features to identify an author, struggle to distinguish between human-authored texts and those generated by LLMs, which adeptly replicate these features.”
The paper relies on this without support.
The diversity in datasets and evaluation metrics currently hinders the comparability and generalizability of different authorship attribution methods.
“The diversity in datasets and evaluation metrics currently hinders the comparability and generalizability of different authorship attribution methods.”
The paper relies on this without support.
Authorship attribution techniques are essential in digital forensics, cybersecurity, and plagiarism detection.
“Authorship attribution techniques are essential in digital forensics, cybersecurity, and plagiarism detection.”
The paper relies on this without support.
The attribution of text to specific authors or models raises significant ethical and privacy issues.
“The attribution of text to specific authors or models raises significant ethical and privacy issues.”
The paper relies on this without support.
The potential to reveal the identities of anonymous authors presents significant ethical challenges.
“However, the potential to reveal the identities of anonymous authors presents significant ethical challenges.”
The paper relies on this without support.
Applications such as linking user accounts across platforms and identifying compromised accounts raise privacy concerns and ethical questions about surveillance and profiling individuals based on their writing style.
“Applications such as linking user accounts across platforms and identifying compromised accounts raise privacy concerns and ethical questions about surveillance and profiling individuals based on their writing style.”
The paper relies on this without support.
Despite existing measures to prohibit the unethical use of LLMs, these restrictions could be evaded through prompt engineering and jail-breaking, posing risks of phishing and fraud scams.
“Despite existing measures to prohibit the unethical use of LLMs, these restrictions could be evaded through prompt engineering and jail-breaking, posing risks of phishing and fraud scams.”
The paper relies on this without support.
The increasing difficulty in distinguishing between human and LLM-generated content raises concerns about intellectual property, plagiarism, and accountability.
“Furthermore, the increasing difficulty in distinguishing between human and LLM-generated content raises concerns about intellectual property, plagiarism, and accountability.”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.3 Open Challenges
“In contrast, LLMgenerated texts typically exhibit consistent syntax and a broader vocabulary [Guo et al., 2023].”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 5.2 Methodologies
“Human-authored texts tend to be more coherent and exhibit greater lexical diversity compared to LLM-generated texts [Guo et al., 2023;Dugan et al., 2023;Zhang et al., 2024].”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > 6.1 Benchmarks and Datasets
“• Aston 100 Idiolects Corpus [Heini and Kredens, 2021]: Comprising emails, essays, text messages, and business memos from 100 individuals (ages 18-22, native English speakers), this corpus provides a broad spectrum of text types for analyzing both content and stylistic features.”
§ Authorship Attribution in the Era of LLMs: Problems, Methodologies, and Challenges > References
“78. Pezik-P. Heini, A. and K. Kredens. 2021. GPTZero: To- wards detection of AI-generated text using zero-shot and supervised methods. http://fold.aston.ac.uk/handle/ 123456789/17”
Citations
Cites
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 1×
“Pre-trained language models, especially BERT-based architectures (Devlin et al., 2019) such as BERT (Fabien et al., 2020, Ippolito et al., 2020, Manolache et al., 2021), Sentence-BERT (Rivera-Soto et al., 2021, Schlicht and de Paula, 2021), and RoBERTa (Huertas-Tato et al., 2022), have proven…”
A Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372 · cited 3×
“The increasingly large context length of LLMs enables in-context learning (ICL) (Brown et al., 2020) by incorporating more documents, enhancing the model’s ability to capture intricate writing patterns.”
InfluentialRoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515 · cited 2×
“Neural network-based detectors, particularly those utilizing BERT, have proven effective in distinguishing between human-written texts and those generated by GPT-2 (Ippolito et al., 2020, Liu et al., 2019).”
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929 · cited 2×
“Models like DNA-GPT (Yang et al., 2023a) and DetectGPT (Mitchell et al., 2023) utilize the T5 model (Raffel et al., 2020) to simulate scenarios where humans modify LLM-generated texts.”
Rudin, C. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0048-x · Cited by 10,561 · cited 1×
“The challenge remains to provide higher-level explainability that aligns with human cognitive processes (Rudin, 2019).”
Explaining and Harnessing Adversarial Examples
Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014) · arXiv · doi:10.48550/arxiv.1412.6572 · Cited by 10,381 · cited 1×
“Neural network-based detectors are vulnerable to adversarial and poisoning attacks (Goodfellow et al., 2015, Pu et al., 2023a, Wang et al., 2022).”
Domain-Adversarial Training of Neural Networks
Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., Marchand, M., & Lempitsky, V. (2017) · Advances in Computer Vision and Pattern Recognition · doi:10.1007/978-3-319-58347-1_10 · Cited by 8,318 · cited 1×
“…n-grams (Madigan et al., 2005b, Menon and Choi, 2011), while recent approaches highlight the importance of training more powerful transformer-based models (Rivera-Soto et al., 2021) and techniques such as adversarial training (Ben-David et al., 2010, Ganin et al., 2016, Li et al., 2017).”
Bag of Tricks for Efficient Text Classification
Joulin, A., Grave, E., Bojanowski, P., & Mikolov, T. (2017) · Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers · doi:10.18653/v1/e17-2068 · Cited by 4,545 · cited 1×
“…Information Density (UID) (Venkatraman et al., 2023), log rank information (Su et al., 2023b), and various linguistic features such as part-of-speech deter-miners, conjunctions, auxiliary relations, vocabulary, and emotional tone (Gehrmann et al., 2019, Joulin et al., 2017, Tang et al., 2023).”
Universal Language Model Fine-tuning for Text Classification
Howard, J., & Ruder, S. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p18-1031 · Cited by 4,100 · cited 1×
“Techniques like slanted triangular learning rates and gradual unfreezing can be used to avoid catastrophic forgetting during fine-tuning (Howard and Ruder, 2018).”
Supervised Contrastive Learning
Khosla, P., Teterwak, P., Wang, C., Sarna, A., Tian, Y., Isola, P., Maschinot, A., Liu, C., & Krishnan, D. (2020) · arXiv · doi:10.48550/arxiv.2004.11362 · Cited by 3,569 · cited 1×
“Contrastive learning (Khosla et al., 2020) is often used with pre-trained language models to enhance stylistic representation by maximizing similarity between texts written by the same authors and minimizing it with texts from different authors (Huertas-Tato et al., 2022).”
A theory of learning from different domains
Ben-David, S., Blitzer, J., Crammer, K., Kulesza, A., Pereira, F., & Vaughan, J. W. (2010) · Machine Learning · doi:10.1007/s10994-009-5152-4 · Cited by 3,515 · cited 1×
“…n-grams (Madigan et al., 2005b, Menon and Choi, 2011), while recent approaches highlight the importance of training more powerful transformer-based models (Rivera-Soto et al., 2021) and techniques such as adversarial training (Ben-David et al., 2010, Ganin et al., 2016, Li et al., 2017).”
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-T., Rocktäschel, T., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.11401 · Cited by 3,101
Governors’ Resilience Tweeting and Community Compliance during the COVID-19 Pandemic
Pennebaker, J., Boyd, R. L., Jordan, K., Blackburn, K. G., Argamon, S., Koppel, M., Pennebaker, J., Schler, J., Bäck, M., Kfner, A., et al. (2021) · Volume 48, Number 6, December 2021 · doi:10.1287/orms.2021.06.14m · Cited by 2,488 · cited 1×
“…(LIWC) tool exemplifies how automated text analysis can use more than 100 psychological dimensions to analyze word use, reflecting distinct language variations among different groups in specific genres and languages (Dudău and Sava, 2021, Goldstein-Stewart et al., 2009, Pennebaker et al., 2015).”
Chao, Y. R. (1950) · Language · doi:10.2307/409735 · Cited by 2,210 · cited 1×
“…been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law (Piantadosi, 2014, Zipf, 2016) 2 .”
Large language models are zero-shot reasoners
Kojima, T., Gu, S. (., Reid, M., Matsuo, Y., & Iwasawa, Y. (2022) · Advances in Neural Information Processing Systems 35 · doi:10.52202/068431-1613 · Cited by 2,013 · cited 1×
“Authorship attribution is a complex reasoning task, and LLMs possess significant capabilities in reasoning and problem-solving, particularly in zero-shot learning within resource-limited domains (Kojima et al., 2022).”
Jindal, N., & Liu, B. (2008) · Proceedings of the international conference on Web search and web data mining - WSDM '08 · doi:10.1145/1341531.1341560 · Cited by 1,542 · cited 1×
“…misinformation and disinformation on social media (Chen and Shu, 2024b, Gambini et al., 2022, Stiff and Johansson, 2022), identifying spam (Jindal and Liu, 2008), preventing phishing attacks (Hazell, 2023), identifying fake reviews (Salminen et al., 2022), detecting machine-generated…”
Justifying Recommendations using Distantly-Labeled Reviews and Fine-Grained Aspects
Ni, J., Li, J., & McAuley, J. (2019) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d19-1018 · Cited by 1,400 · cited 2×
“Therefore, the representative datasets for human authorship are summarized as follows: • Amazon Review (Ni et al., 2019): Featuring reviews with ratings, text, votes, product metadata, and links, this dataset provides a comprehensive view of consumer opinions, ideal for commercial authorship…”
The Curious Case of Neural Text Degeneration
Holtzman, A., Buys, J., Du, L., Forbes, M., & Choi, Y. (2019) · arXiv · doi:10.48550/arxiv.1904.09751 · Cited by 1,304 · cited 1×
“…been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law (Piantadosi, 2014, Zipf, 2016) 2 .”
The handbook of language variation and change
Chambers, J., Trudgill, P., & Schilling-Estes, N. (2003) · doi:10.1111/b.9781405116923.2003.x · Cited by 1,262 · cited 1×
“…markers distinctively from lexical content words, revealing correlations with socio-cultural categories such as gender, age, and native language (Chambers et al., 2013, Nerbonne, 2014, Seals and Shalin, 2023), which are pivotal in understanding identity and social affiliations (Argamon et al.,…”
The Enron Corpus: A New Dataset for Email Classification Research
Klimt, B., & Yang, Y. (2004) · Lecture notes in computer science · doi:10.1007/978-3-540-30115-8_22 · Cited by 1,238 · cited 1×
“This holistic approach not only improves the attribution process but also fosters the development of innovative methodologies that are better adapted to the evolving nature of digital content.”
Narayan, S., Cohen, S. B., & Lapata, M. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1206 · Cited by 1,207
Baumgartner, J., Zannettou, S., Keegan, B., Squire, M., & Blackburn, J. (2020) · International Conference on Web and Social Media · doi:10.5281/zenodo.3608135 · Cited by 1,178 · cited 1×
“• Pushshift Reddit (Baumgartner et al., 2020): This dataset comprises posts and comments from various subreddits, covering diverse topics and writing styles, making it suitable for analyzing informal online discourse.”
A survey of modern authorship attribution methods
Stamatatos, E. (2009) · Journal of the American Society for Information Science and Technology · doi:10.1002/asi.21001 · Cited by 1,176 · cited 1×
“Authorship attribution is inherently multidisciplinary, encompassing elements of linguistics, computer science, forensic science, and psychology (Stamatatos, 2009).”
Gao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2022) · doi:10.1101/2022.12.23.521610 · Cited by 1,099 · cited 1×
“Identifying LLM-generated texts is challenging even for human experts, let alone traditional authorship attribution methods (Gao et al., 2022, Liu et al., 2023a).”
Finding Deceptive Opinion Spam by Any Stretch of the Imagination
Ott, M., Choi, Y., Cardie, C., & Hancock, J. T. (2011) · arXiv · doi:10.48550/arxiv.1107.4557 · Cited by 1,008 · cited 2×
“• Deceptive Opinion Spam (Ott et al., 2011): Includes 400 genuine and 400 deceptive hotel reviews, with deceptive reviews generated using Amazon Mechanical Turk, useful for studying the nuances of fake versus real reviews.”
Zipf’s word frequency law in natural language: A critical review and future directions
Piantadosi, S. T. (2014) · Psychonomic Bulletin & Review · doi:10.3758/s13423-014-0585-6 · Cited by 955 · cited 1×
“…been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law (Piantadosi, 2014, Zipf, 2016) 2 .”
Lund, B. D., Wang, T., Mannuru, N. R., Nie, B., Shimray, S., & Wang, Z. (2023) · Journal of the Association for Information Science and Technology · doi:10.1002/asi.24750 · Cited by 919 · cited 2×
“Mis-attribution can lead to wrongful accusations or misinterpretation of an author’s intent (Lund et al., 2023).”
Baumgartner, J., Zannettou, S., Keegan, B., Squire, M., & Blackburn, J. (2020) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v14i1.7347 · Cited by 881
Can AI-Generated Text be Reliably Detected?
Sadasivan, V. S., Kumar, A., Balasubramanian, S., Wang, W., & Feizi, S. (2023) · arXiv · doi:10.48550/arxiv.2303.11156 · Cited by 810 · cited 4×
“For instance, a 10-class classification might include seven human classes and three LLM classes (Tripto et al., 2023).”
InfluentialGPT detectors are biased against non-native English writers
Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023) · Patterns · doi:10.1016/j.patter.2023.100779 · Cited by 673 · cited 2×
“…their effectiveness decreases as the number of candidate authors increases due to context length constraints (Huang et al., 2024), and they can perpetuate biases from training data, resulting in inaccurate attributions for texts from marginalized groups and languages (Liang et al., 2023).”
Defending Against Neural Fake News
Zellers, R., Holtzman, A., Rashkin, H., Bisk, Y., Farhadi, A., Roesner, F., & Choi, Y. (2019) · arXiv · doi:10.48550/arxiv.1905.12616 · Cited by 569 · cited 4×
“Typically, this task is approached as a binary classification problem (Fagni et al., 2021, Jawahar et al., 2020, Mitchell et al., 2023, Solaiman et al., 2019, Zellers et al., 2019).”
InfluentialGender, genre, and writing style in formal written texts
ARGAMON, S., KOPPEL, M., FINE, J., & SHIMONI, A. R. (2003) · Text - Interdisciplinary Journal for the Study of Discourse · doi:10.1515/text.2003.014 · Cited by 560 · cited 1×
“Second, developing domain and genre-invariant features would improve robustness by focusing on core stylistic elements (Argamon et al., 2003).”
Explain Yourself! Leveraging Language Models for Commonsense Reasoning
Rajani, N. F., McCann, B., Xiong, C., & Socher, R. (2019) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p19-1487 · Cited by 555 · cited 1×
“Advances in this area may include leveraging discourse-level relations and training models with human explanations for common sense reasoning to improve the explanatory depth of model-generated attributions (Rajani et al., 2019).”
Juola, P. (2008) · Foundations and Trends® in Information Retrieval · doi:10.1561/1500000005 · Cited by 530 · cited 1×
“…and frequency; syntactic features pertain to sentence structure and grammar; semantic features explore the meaning and context of words; structural features relate to text organization; and content-specific features emphasize domain-specific terms (Bozkurt et al., 2007, Seroussi et al., 2014).”
GLTR: Statistical Detection and Visualization of Generated Text
Gehrmann, S., Strobelt, H., & Rush, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations · doi:10.18653/v1/p19-3019 · Cited by 470 · cited 5×
“Various statistical measures have been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law…”
InfluentialA Survey of the State of Explainable AI for Natural Language Processing
Danilevsky, M., Qian, K., Aharonov, R., Katsis, Y., Kawas, B., & Sen, P. (2020) · Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the International Joint Conference on Natural Language Processing (AACL-IJCNLP) · doi:10.18653/v1/2020.aacl-main.46 · Cited by 468 · cited 1×
“This challenge underscores the need for enhanced methodologies that not only differentiate between human and AI-generated content but also explore how LLMs emulate specific authorial styles (Boenninghoff et al., 2019a, Danilevsky et al., 2020).”
Mosteller, F., & Wallace, D. L. (1963) · Journal of the American Statistical Association · doi:10.1080/01621459.1963.10500849 · Cited by 466
Automatically profiling the author of an anonymous text
Argamon, S., Koppel, M., Pennebaker, J. W., & Schler, J. (2009) · Communications of the ACM · doi:10.1145/1461928.1461959 · Cited by 376 · cited 4×
“This discipline utilizes a variety of linguistic features to determine authorship (Holmes, 1994, Lagutina et al., 2019), positing that each author’s unique style can be captured through quantifiable characteristics (Argamon et al., 2009).”
InfluentialCreating and detecting fake reviews of online products
Salminen, J. O., Kandpal, C. S., Kamel, A. M., Jung, S., & Jansen, B. J. J. (2021) · Journal of Retailing and Consumer Services · doi:10.1016/j.jretconser.2021.102771 · Cited by 362 · cited 1×
“…2022, Stiff and Johansson, 2022), identifying spam (Jindal and Liu, 2008), preventing phishing attacks (Hazell, 2023), identifying fake reviews (Salminen et al., 2022), detecting machine-generated scientific papers (Liu et al., 2023a, Rodriguez et al., 2022), and making academic peer review…”
DetectLLM: Leveraging Log Rank Information for Zero-Shot Detection of Machine-Generated TextPreprint
Su, J., Zhuo, T. Y., Wang, D., & Nakov, P. (2023) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2306.05540 · Cited by 347
RADAR: Robust AI-Text Detection via Adversarial LearningPreprint
Hu, X., Chen, P.-Y., & Ho, T.-Y. (2023) · Neural Information Processing Systems · doi:10.48550/arxiv.2307.03838 · Cited by 328 · cited 2×
“There are fewer studies on this task due to its difficulty, and existing research often makes simplifications.”
Detecting Hoaxes, Frauds, and Deception in Writing Style Online
Afroz, S., Brennan, M., & Greenstadt, R. (2012) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp.2012.34 · Cited by 320 · cited 1×
“…(Chen and Shu, 2024a,b, Hanley and Durumeric, 2024, Shu et al., 2020, Stiff and Johansson, 2022), protecting intellectual property rights (Meyer zu Eissen et al., 2007, Stamatatos and Koppel, 2011), and identifying fraudulent practices such as fake reviews (Afroz et al., 2012, Ott et al., 2011).”
How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection
Guo, B., Zhang, X., Wang, Z., Jiang, M., Nie, J., Ding, Y., Yue, J., & Wu, Y. (2023) · arXiv · doi:10.48550/arxiv.2301.07597 · Cited by 318 · cited 13×
“Recently, many commercial platforms have also integrated “Humanizer” functions, reflecting an emerging detector-evasion ecosystem in which detection and text rewriting tools co-evolve.”
InfluentialThe State of Authorship Attribution Studies: Some Problems and Solutions
Rudman, J. (1997) · Computers and the Humanities · doi:10.1023/a:1001018624850 · Cited by 300 · cited 1×
“Important features can be categorized into the following types: lexical, syntactic, semantic, structural, and content-specific (Rudman, 1997).”
Release Strategies and the Social Impacts of Language Models
Solaiman, I., Brundage, M., Clark, J., Askell, A., Herbert-Voss, A., Wu, J., Radford, A., Krueger, G., Kim, J. W., Kreps, S., et al. (2019) · arXiv · doi:10.48550/arxiv.1908.09203 · Cited by 292 · cited 6×
“Solaiman et al. (2019) fine-tuned the RoBERTa model with a dataset of GPT-2 outputs in open domain settings.”
InfluentialAll That’s ‘Human’ Is Not Gold: Evaluating Human Evaluation of Generated Text
Clark, E., August, T., Serrano, S., Haduong, N., Gururangan, S., & Smith, N. A. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.565 · Cited by 290 · cited 2×
“The advent of LLMs has complicated authorship attribution, making it increasingly difficult to distinguish between LLM-generated texts and human-written texts (Clark et al., 2021, Sadasivan et al., 2023).”
Holmes, D. I. (1994) · Computers and the Humanities · doi:10.1007/bf01830689 · Cited by 272
User Identity Linkage across Online Social Networks
Shu, K., Wang, S., Tang, J., Zafarani, R., & Liu, H. (2017) · ACM SIGKDD Explorations Newsletter · doi:10.1145/3068777.3068781 · Cited by 270 · cited 1×
“…digital realm, authorship attribution helps safe-guard the integrity of content by preventing deceptive social media activities (Hazell, 2023), detecting account compromises (Barbon et al., 2017), and linking user profiles across various social networks (Shu et al., 2017, Sinnott and Wang, 2021).”
Self-citationAutomatic Detection of Generated Text is Easiest when Humans are Fooled
Ippolito, D., Duckworth, D., Callison-Burch, C., & Eck, D. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.164 · Cited by 264 · cited 17×
“For instance, Kowalczyk et al. (2022) detected GPT-2-generated fake reviews using Shapley Additive Explanations (SHAP) (Lundberg and Lee, 2017).”
InfluentialEvaluating the Social Impact of Generative AI Systems in Systems and SocietyPreprint
Solaiman, I., Talat, Z., Agnew, W. S., Ahmad, L., Baker, D., Blodgett, S. L., Chen, C., Daumé, H., Dodge, J., Duan, I., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2306.05949 · Cited by 255
Self-citationAI model GPT-3 (dis)informs us better than humans
Spitale, G., Biller-Andorno, N., & Germani, F. (2023) · Science Advances · doi:10.1126/sciadv.adh1850 · Cited by 241 · cited 1×
“Future research should focus on refining existing authorship attri-bution methods to detect and prevent malicious activities such as generating misinformation, plagiarism, and propaganda (Goldstein et al., 2023, Hazell, 2023, Lund et al., 2023, Spitale et al., 2023).”
End-to-End Adversarial Memory Network for Cross-domain Sentiment Classification
Li, Z., Zhang, Y., Wei, Y., Wu, Y., & Yang, Q. (2017) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2017/311 · Cited by 230 · cited 1×
“…n-grams (Madigan et al., 2005b, Menon and Choi, 2011), while recent approaches highlight the importance of training more powerful transformer-based models (Rivera-Soto et al., 2021) and techniques such as adversarial training (Ben-David et al., 2010, Ganin et al., 2016, Li et al., 2017).”
Bigrams of Syntactic Labels for Authorship Discrimination of Short Texts
Hirst, G., & Feiguina, O. (2007) · Literary and Linguistic Computing · doi:10.1093/llc/fqm023 · Cited by 227 · cited 1×
“…the performance of existing authorship attri-bution models include the size of the training text (Hirst and Feigu-ina, 2007, Marton et al., 2005), the number of candidate authors (Koppel et al., 2006), and the imbalanced distribution of training texts among the candidate authors (Stamatatos, 2008).”
Can LLM-Generated Misinformation Be Detected?Preprint
Chen, C., & Shu, K. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.13788 · Cited by 224 · cited 1×
“Additionally, authorship attribution techniques are instrumental in combating misinformation (Chen and Shu, 2024a,b, Hanley and Durumeric, 2024, Shu et al., 2020, Stiff and Johansson, 2022), protecting intellectual property rights (Meyer zu Eissen et al., 2007, Stamatatos and Koppel, 2011), and…”
Self-citationSurveying Stylometry Techniques and Applications
Neal, T., Sundararajan, K., Fatima, A., Yan, Y., Xiang, Y., & Woodard, D. (2018) · ACM Computing Surveys · doi:10.1145/3132039 · Cited by 222 · cited 1×
“Stylometry, the quantitative analysis of writing style, has evolved from its initial reliance on human expertise (Mosteller and Wallace, 1963) to computational methods (Lagutina et al., 2019, Neal et al., 2017).”
Authorship Attribution for Social Media Forensics
Rocha, A., Scheirer, W. J., Forstall, C. W., Cavalcante, T., Theophilo, A., Shen, B., Carvalho, A. R. B., & Stamatatos, E. (2017) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2016.2603960 · Cited by 222
Does Size Matter? Authorship Attribution, Small Samples, Big Problem
Eder, M. (2015) · Digital Scholarship in the Humanities · doi:10.1093/llc/fqt066 · Cited by 214 · cited 4×
“Short or low-quality texts, such as social media posts, often provide unreliable data, complicating accurate attribution (Eder, 2015, Theophilo et al., 2021).”
InfluentialA Survey on LLM-Generated Text Detection: Necessity, Methods, and Future Directions
Wu, J., Yang, S., Zhan, R., Yuan, Y., Wong, D. F., & Chao, L. S. (2023) · arXiv · doi:10.48550/arxiv.2310.14724 · Cited by 212
Best practices for the human evaluation of automatically generated text
van der Lee, C., Gatt, A., van Miltenburg, E., Wubben, S., & Krahmer, E. (2019) · Proceedings of the 12th International Conference on Natural Language Generation · doi:10.18653/v1/w19-8643 · Cited by 205 · cited 1×
“Evaluating the quality of machine-generated excerpts has traditionally relied on human judgment, which is considered the gold standard for open-domain generation systems (Gehrmann et al., 2019, van der Lee et al., 2019).”
Natural Language Watermarking: Design, Analysis, and a Proof-of-Concept Implementation
Atallah, M. J., Raskin, V., Crogan, M., Hempelmann, C. F., Kerschbaum, F., Mohamed, D., & Naik, S. (2001) · Lecture notes in computer science · doi:10.1007/3-540-45496-9_14 · Cited by 201 · cited 1×
“Various methods include parsed syntactic tree structures (Atallah et al., 2001, Topkara et al., 2005), synonym tables (Jalil and Mirza, 2009), adversarial watermarking (Abdelnabi and Fritz, 2021), and context-aware lexical substitution (Yang et al., 2022).”
Machine-Generated Text: A Comprehensive Survey of Threat Models and Detection Methods
Crothers, E. N., Japkowicz, N., & Viktor, H. L. (2023) · IEEE Access · doi:10.1109/access.2023.3294090 · Cited by 196
A Watermark for Large Language Models
Kirchenbauer, J., Geiping, J., Wen, Y., Katz, J., Miers, I., & Goldstein, T. (2023) · arXiv · doi:10.48550/arxiv.2301.10226 · Cited by 194
Bozkurt, İ. N., Baghoglu, O., & Uyar, E. (2007) · International Symposium on Computer and Information Sciences · doi:10.1109/iscis.2007.4456854 · Cited by 187
TweepFake: about detecting deepfake tweets
Fagni, T., Falchi, F., Gambini, M., Martella, A., & Tesconi, M. (2021) · PLOS ONE · doi:10.1371/journal.pone.0251415 · Cited by 185 · cited 1×
“Typically, this task is approached as a binary classification problem (Fagni et al., 2021, Jawahar et al., 2020, Mitchell et al., 2023, Solaiman et al., 2019, Zellers et al., 2019).”
Convolutional Neural Networks for Authorship Attribution of Short Texts
Shrestha, P., Sierra, S., Gonzalez, F. A., Montes, M., Rosso, P., & Solorio, T. (2017) · Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/e17-2106 · Cited by 172 · cited 1×
“Convolutional Neural Networks (CNNs) were applied at various levels, including characters, words, and N-grams (Ruder et al., 2016, Shrestha et al., 2017).”
Sison, A. J. G., Daza, M. T., Gozalo-Brizuela, R., & Garrido-Merchán, E. C. (2024) · International Journal of Human–Computer Interaction · doi:10.1080/10447318.2023.2225931 · Cited by 169
Determining an author's native language by mining a text for errors
Koppel, M., Schler, J., & Zigdon, K. (2005) · Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining · doi:10.1145/1081870.1081947 · Cited by 167 · cited 1×
“Current authorship attri-bution models also struggle with out-of-distribution issues when faced with languages and LLMs not encountered during training, leading to decreased accuracy and reliability (Koppel et al., 2005, Wu et al., 2023b).”
Automatic Detection of Machine Generated Text: A Critical Survey
Jawahar, G., Abdul-Mageed, M., & Lakshmanan, V. S. (2020) · Proceedings of the 28th International Conference on Computational Linguistics · doi:10.18653/v1/2020.coling-main.208 · Cited by 156 · cited 3×
“Typically, this task is approached as a binary classification problem (Fagni et al., 2021, Jawahar et al., 2020, Mitchell et al., 2023, Solaiman et al., 2019, Zellers et al., 2019).”
Goldstein, J. A., Sastry, G., Musser, M., DiResta, R., Gentzel, M., & Sedova, K. (2023) · arXiv · doi:10.48550/arxiv.2301.04246 · Cited by 152
DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvature
Mitchell, E., Lee, Y., Khazatsky, A., Manning, C. D., & Finn, C. (2023) · arXiv · doi:10.48550/arxiv.2301.11305 · Cited by 151 · cited 15×
“These detectors require retraining when encountering text from new LLMs to ensure reliable detection (Chakraborty et al., 2023b, Mitchell et al., 2023).”
InfluentialParaphrasing evades detectors of AI-generated text, but retrieval is an effective defense
Krishna, K., Song, Y., Karpinska, M., Wieting, J., & Iyyer, M. (2023) · arXiv · doi:10.48550/arxiv.2303.13408 · Cited by 150 · cited 2×
“Additionally, Krishna et al. (2024) developed a detector that uses information retrieval to store LLM outputs in a database and search for semantically similar content to identify LLM-generated text, though this method raises privacy concerns regarding the storage of user conversations.”
Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense
Krishna, K., Song, Y., Karpinska, M., Wieting, J., & Iyyer, M. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-1195 · Cited by 150
Combating misinformation in the age of LLMs: Opportunities and challenges
Chen, C., & Shu, K. (2024) · AI Magazine · doi:10.1002/aaai.12188 · Cited by 148
Self-citationMGTBench: Benchmarking Machine-Generated Text Detection
He, X., Shen, X., Chen, Z., Michael Backes, A., & Zhang, Y. (2024) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3658644.3670344 · Cited by 142 · cited 17×
“Neural network-based detectors, generally outperform metric-based methods in both human authorship attribution and LLM-generated text detection problems (He et al., 2023, Zhang et al., 2024).”
InfluentialDo Massively Pretrained Language Models Make Better Storytellers?
See, A., Pappu, A., Saxena, R., Yerukola, A., & Manning, C. D. (2019) · Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL) · doi:10.18653/v1/k19-1079 · Cited by 139 · cited 1×
“Similar to the problem of human authorship attribution, linguistic features such as phrasal verbs, coreference, part-of-speech (POS) tags, and named entity (NE) tags are also useful in distinguishing LLM-generated text (Fröhling and Zubiaga, 2021, Nguyen-Son et al., 2017, See et al., 2019).”
Dugan, L., Ippolito, D., Kirubarajan, A., et al. (2022) · AAAI Conference on Artificial Intelligence · doi:10.48550/arxiv.2212.12672 · Cited by 138 · cited 8×
“Table 3: Overview of open-source LLM-generated text detectors. mainly designed for LLM-generated text detection (Problem 2), while some, such as GPTZero (Tian and Cui, 2023), also support human-LLM co-authorship analysis and boundary detection, partially addressing Problem 4 (Cutler et al., 2021).”
InfluentialAuthor identification: Using text sampling to handle the class imbalance problem
Stamatatos, E. (2007) · Information Processing & Management · doi:10.1016/j.ipm.2007.05.012 · Cited by 136 · cited 1×
“…the performance of existing authorship attri-bution models include the size of the training text (Hirst and Feigu-ina, 2007, Marton et al., 2005), the number of candidate authors (Koppel et al., 2006), and the imbalanced distribution of training texts among the candidate authors (Stamatatos, 2008).”
Mining Disinformation and Fake News: Concepts, Methods, and Recent Advancements
Shu, K., Wang, S., Lee, D., & Liu, H. (2020) · Lecture notes in social networks · doi:10.1007/978-3-030-42699-6_1 · Cited by 124 · cited 1×
“Additionally, authorship attribution techniques are instrumental in combating misinformation (Chen and Shu, 2024a,b, Hanley and Durumeric, 2024, Shu et al., 2020, Stiff and Johansson, 2022), protecting intellectual property rights (Meyer zu Eissen et al., 2007, Stamatatos and Koppel, 2011), and…”
Self-citationAuthorship Attribution for Neural Text Generation
Uchendu, A., Le, T., Shu, K., & Lee, D. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.673 · Cited by 120 · cited 2×
“These detectors also face limitations such as overfitting to training data (Uchendu et al., 2020).”
Self-citationAdversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hiding
Abdelnabi, S., & Fritz, M. (2021) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp40001.2021.00083 · Cited by 120 · cited 2×
“MIXSET (Zhang et al., 2024) offers a more comprehensive dataset that includes text refined by LLMs through polishing, completion, and rewriting operations.”
AllenNLP Interpret: A Framework for Explaining Predictions of NLP Models
Wallace, E., Tuyls, J., Wang, J., Subramanian, S., Gardner, M., & Singh, S. (2019) · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP): System Demonstrations · doi:10.18653/v1/d19-3002 · Cited by 116
SemStamp: A Semantic Watermark with Paraphrastic Robustness for Text Generation
Hou, A., Zhang, J., He, T., Wang, Y., Chuang, Y.-S., Wang, H., Shen, L., Durme, B. V., Khashabi, D., & Tsvetkov, Y. (2024) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2024.naacl-long.226 · Cited by 115 · cited 1×
“Further advancements have improved the robustness, efficiency, and stealthiness of watermarking methods (Hou et al., 2023, Wu et al., 2023a, Zhao et al., 2023).”
Natural language watermarking via morphosyntactic alterations
Meral, H. M., Sankur, B., Sumru Özsoy, A., Güngör, T., & Sevinç, E. (2009) · Computer Speech & Language · doi:10.1016/j.csl.2008.04.001 · Cited by 113 · cited 1×
“(Kirchenbauer et al., 2023, Meral et al., 2009, Topkara et al., 2005, Zhao et al., 2023).”
The limitations of stylometry for detecting machine-generated fake news
Schuster, T., Schuster, R., Shah, D. J., & Barzilay, R. (2020) · Computational Linguistics · doi:10.1162/coli_a_00380 · Cited by 109 · cited 2×
“One variation of this task is to detect the boundary between human-written and LLM-generated text (Cutler et al., 2021, Wang et al., 2024).”
A Survey on Detection of LLMs-Generated ContentPreprint
Yang, X., Pan, L., Zhao, X., Chen, H.-F., Petzold, L., Wang, W. Y., & Cheng, W. (2023) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2310.15654 · Cited by 108 · cited 3×
“Variations in model architecture (such as the number of layers and parameters), training methods (including pre-training and fine-tuning), and generation techniques (like sampling parameters) all subtly influence the characteristics of the generated texts (Munir et al., 2021).”
Plagiarism Detection Without Reference Collections
Meyer zu Eissen, S., Stein, B., & Kulig, M. (2007) · Studies in Classification, Data Analysis, and Knowledge Organization · doi:10.1007/978-3-540-70981-7_40 · Cited by 106 · cited 2×
“…(Chen and Shu, 2024a,b, Hanley and Durumeric, 2024, Shu et al., 2020, Stiff and Johansson, 2022), protecting intellectual property rights (Meyer zu Eissen et al., 2007, Stamatatos and Koppel, 2011), and identifying fraudulent practices such as fake reviews (Afroz et al., 2012, Ott et…”
Feature-based detection of automated language models: tackling GPT-2, GPT-3 and Grover
Fröhling, L., & Zubiaga, A. (2021) · PeerJ Computer Science · doi:10.7717/peerj-cs.443 · Cited by 106
The Secret Life of Pronouns. What Our Words Say About Us
Nerbonne, J. (2014) · Literary and Linguistic Computing · doi:10.1093/llc/fqt006 · Cited by 93 · cited 1×
“…distinctively from lexical content words, revealing correlations with socio-cultural categories such as gender, age, and native language (Chambers et al., 2013, Nerbonne, 2014, Seals and Shalin, 2023), which are pivotal in understanding identity and social affiliations (Argamon et al., 2009).”
Authorship attribution with thousands of candidate authors
Koppel, M., Schler, J., Argamon, S., & Messeri, E. (2006) · Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval · doi:10.1145/1148170.1148304 · Cited by 90
Authorship Attribution with Topic Models
Seroussi, Y., Zukerman, I., & Bohnert, F. (2014) · Computational Linguistics · doi:10.1162/coli_a_00173 · Cited by 89 · cited 4×
“…and frequency; syntactic features pertain to sentence structure and grammar; semantic features explore the meaning and context of words; structural features relate to text organization; and content-specific features emphasize domain-specific terms (Bozkurt et al., 2007, Seroussi et al., 2014).”
InfluentialCHEAT: A Large-Scale Dataset for Detecting CHatGPT-writtEn AbsTracts
Yu, P., Chen, J., Feng, X., & Xia, Z. (2025) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2025.3536929 · Cited by 88
Ruder, S., Ghaffari, P., & Breslin, J. G. (2016) · arXiv · doi:10.48550/arxiv.1609.06686 · Cited by 86 · cited 1×
“Convolutional Neural Networks (CNNs) were applied at various levels, including characters, words, and N-grams (Ruder et al., 2016, Shrestha et al., 2017).”
TURINGBENCH: A Benchmark Environment for Turing Test in the Age of Neural Text Generation
Uchendu, A., Ma, Z., Le, T., Zhang, R., & Lee, D. (2021) · Findings of the Association for Computational Linguistics: EMNLP 2021 · doi:10.18653/v1/2021.findings-emnlp.172 · Cited by 82 · cited 6×
“To simplify the classification process, it is common practice to group different human writers into a single category because humans exhibit a broader spectrum of writing styles and proficiency levels compared to machines (Uchendu et al., 2021).”
InfluentialA Review of Digital Watermarking Techniques for Text Documents
Jalil, Z., & Mirza, A. M. (2009) · International Conference on Information and Multimedia Technology · doi:10.1109/icimt.2009.11 · Cited by 81 · cited 1×
“Various methods include parsed syntactic tree structures (Atallah et al., 2001, Topkara et al., 2005), synonym tables (Jalil and Mirza, 2009), adversarial watermarking (Abdelnabi and Fritz, 2021), and context-aware lexical substitution (Yang et al., 2022).”
Author Identification Using Multi-headed Recurrent Neural Networks
Bagnall, D. (2015) · arXiv · doi:10.48550/arxiv.1506.04891 · Cited by 80 · cited 1×
“Before the widespread adoption of transformer-based models, multi-headed Recurrent Neural Networks (RNNs) (Bagnall, 2015), and Long Short-Term Memory (LSTMs) were utilized at both sentence and article levels (Qian et al., 2017).”
Quantifying evidence in forensic authorship analysis
Grant, T. (2007) · International Journal of Speech Language and the Law · doi:10.1558/ijsll.v14i1.1 · Cited by 80 · cited 1×
“Detecting LLM-generated text is crucial for several applications, including combating misinformation and disinformation on social media (Chen and Shu, 2024b, Gambini et al., 2022, Stiff and Johansson, 2022), identifying spam (Jindal and Liu, 2008), preventing phishing attacks (Hazell, 2023),…”
Liu, Z. Q., Yao, Z., Li, F., & Luo, B. (2024) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3658644.3670392 · Cited by 80 · cited 3×
“Identifying LLM-generated texts is challenging even for human experts, let alone traditional authorship attribution methods (Gao et al., 2022, Liu et al., 2023a).”
MAGE: Machine-generated Text Detection in the Wild
Li, Y., Li, Q., Cui, L., Bi, W., Wang, Z., Wang, L., Yang, L., Shi, S., & Zhang, Y. (2024) · Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2024.acl-long.3 · Cited by 79 · cited 1×
“LLM-generated text detectors often struggle to generalize to unseen domains encountered during training (Li et al., 2023a, Pu et al., 2023b, Rodriguez et al., 2022) and tend to perform better on LLMs they were specifically trained on (Chakraborty et al., 2023b, Li et al., 2023a, Pu et al., 2023b).”
Detecting computer-generated disinformation
Stiff, H., & Johansson, F. (2022) · International Journal of Data Science and Analytics · doi:10.1007/s41060-021-00299-5 · Cited by 79
Neural Deepfake Detection with Factual Structure of Text
Zhong, W., Tang, D., Xu, Z., Wang, R., Duan, N., Zhou, M., Wang, J., & Yin, J. P. (2020) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.193 · Cited by 73 · cited 2×
“One variation of this task is to detect the boundary between human-written and LLM-generated text (Cutler et al., 2021, Wang et al., 2024).”
Bayesian Multinomial Logistic Regression for Author Identification
Madigan, D., Genkin, A., Lewis, D. D., et al. (2005) · AIP Conference Proceedings · doi:10.1063/1.2149832 · Cited by 68 · cited 2×
“Machine learning approaches integrate stylometric features with classifiers such as logistic regression (Aborisade and Anwar, 2018, Madigan et al., 2005a), Bayesian multinomial regression (Argamon et al., 2009), and support vector machines (SVM) (Bacciu et al., 2019).”
Explainable Authorship Verification in Social Media via Attention-based Similarity Learning
Boenninghoff, B., Hessler, S., Kolossa, D., & Nickel, R. M. (2019) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata47090.2019.9005650 · Cited by 66 · cited 3×
“Moreover, syntax-augmented CNN models (Zhang et al., 2018), Convolutional Siamese Networks (Saedi and Dras, 2021), and attention-based Siamese Networks (Boenninghoff et al., 2019a) were explored.”
Red Teaming Language Model Detectors with Language Models
Shi, Z., Wang, Y., Yin, F., Chen, X., Chang, K.-W., & Hsieh, C. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00639 · Cited by 65 · cited 1×
“…including homoglyph attacks (Gagiano et al., 2021, Macko et al., 2024), whitespace insertion (Cai and Cui, 2023), syntactic perturbations (Bhat and Parthasarathy, 2020), synonym replacement (Kulkarni et al., 2023), and paraphrasing (Becker et al., 2023, Krishna et al., 2024, Shi et al., 2024).”
Few-Shot Detection of Machine-Generated Text using Style RepresentationsPreprint
Soto, R. A. R., Koch, K., Khan, A., Chen, B., Bishop, M., & Andrews, N. (2024) · International Conference on Learning Representations · doi:10.48550/arxiv.2401.06712 · Cited by 64 · cited 1×
“Manually collecting data for large datasets is time-consuming and costly, motivating researchers to utilize web data sources such as Wikipedia and Red-dit.”
Topkara, M., Taskiran, C. M., & Delp III, E. J. (2005) · SPIE Proceedings · doi:10.1117/12.593790 · Cited by 63 · cited 2×
“(Kirchenbauer et al., 2023, Meral et al., 2009, Topkara et al., 2005, Zhao et al., 2023).”
A Survey on Stylometric Text Features
Lagutina, K., Lagutina, N., Boychuk, E., Vorontsova, I., Shliakhtina, E., Belyaeva, O., Paramonov, I., & Demidov, P. G. (2019) · 2019 25th Conference of Open Innovations Association (FRUCT) · doi:10.23919/fruct48121.2019.8981504 · Cited by 63 · cited 3×
“This discipline utilizes a variety of linguistic features to determine authorship (Holmes, 1994, Lagutina et al., 2019), positing that each author’s unique style can be captured through quantifiable characteristics (Argamon et al., 2009).”
Tracing Text Provenance via Context-Aware Lexical Substitution
Yang, X., Zhang, J., Chen, K., Zhang, W., Ma, Z., Wang, F., & Yu, N. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i10.21415 · Cited by 61 · cited 1×
“Various methods include parsed syntactic tree structures (Atallah et al., 2001, Topkara et al., 2005), synonym tables (Jalil and Mirza, 2009), adversarial watermarking (Abdelnabi and Fritz, 2021), and context-aware lexical substitution (Yang et al., 2022).”
Classification for Authorship of Tweets by Comparing Logistic Regression and Naive Bayes Classifiers
Aborisade, O., & Anwar, M. (2018) · 2018 IEEE International Conference on Information Reuse and Integration (IRI) · doi:10.1109/iri.2018.00049 · Cited by 61 · cited 2×
“Machine learning approaches integrate stylometric features with classifiers such as logistic regression (Aborisade and Anwar, 2018, Madigan et al., 2005a), Bayesian multinomial regression (Argamon et al., 2009), and support vector machines (SVM) (Bacciu et al., 2019).”
On Compression-Based Text Classification
Marton, Y., Wu, N., & Hellerstein, L. (2005) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-31865-1_22 · Cited by 59 · cited 1×
“Factors influencing the performance of existing authorship attri-bution models include the size of the training text (Hirst and Feigu-ina, 2007, Marton et al., 2005), the number of candidate authors (Koppel et al., 2006), and the imbalanced distribution of training texts among the candidate authors…”
On the Possibilities of AI-Generated Text Detection
Chakraborty, S., Bedi, A. S., Zhu, S., An, B., Manocha, D., & Huang, F. (2023) · arXiv · doi:10.48550/arxiv.2304.04736 · Cited by 59
Bevendorff, J., Chulvi, B., Fersini, E., et al. (2022) · Conference and Labs of the Evaluation Forum · doi:10.1007/978-3-031-13643-6_24 · Cited by 55 · cited 1×
“This holistic approach not only improves the attribution process but also fosters the development of innovative methodologies that are better adapted to the evolving nature of digital content.”
Cross-Domain Authorship Attribution Using Pre-trained Language Models
Barlas, G., & Stamatatos, E. (2020) · IFIP advances in information and communication technology · doi:10.1007/978-3-030-49161-1_22 · Cited by 54 · cited 4×
“First, leveraging transfer learning by pre-training on large, diverse datasets and fine-tuning on specific domains enhances adaptability and performance (Barlas and Stamatatos, 2020, Rodriguez et al., 2022).”
InfluentialDugan, L., Ippolito, D., Kirubarajan, A., Shi, S., & Callison-Burch, C. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i11.26501 · Cited by 54
Authorship verification applied to detection of compromised accounts on online social networks
Barbon, S., Igawa, R. A., & Zarpel�ão, B. B. (2016) · Multimedia Tools and Applications · doi:10.1007/s11042-016-3899-8 · Cited by 53 · cited 1×
“…digital realm, authorship attribution helps safe-guard the integrity of content by preventing deceptive social media activities (Hazell, 2023), detecting account compromises (Barbon et al., 2017), and linking user profiles across various social networks (Shu et al., 2017, Sinnott and Wang, 2021).”
The Science of Detecting LLM-Generated TextsPreprint
Tang, R., Chuang, Y.-N., & Hu, X. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2303.07205 · Cited by 53
Syntax Encoding with Application in Authorship Attribution
Zhang, R., Hu, Z., Guo, H., & Mao, Y. (2018) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d18-1294 · Cited by 50 · cited 1×
“As a result, the detection of LLM-generated text has garnered significant attention (Kumarage and Liu, 2023, Tang et al., 2023, Wu et al., 2023b, Yang et al., 2023b).”
Chen, C., Wang, H., Shapiro, M. A., Xiao, Y., Wang, F., & Shu, K. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2211.05289 · Cited by 49
Self-citationOn the State of the Art in Authorship Attribution and Authorship VerificationPreprint
Tyo, J., Dhingra, B., & Lipton, Z. C. (2022) · arXiv.org · doi:10.48550/arxiv.2209.06869 · Cited by 47 · cited 1×
“• VALLA (Tyo et al., 2022): Designed for benchmarking authorship attribution models, VALLA standardizes a range of texts across various genres and writing styles.”
Plagiarism and authorship analysis: introduction to the special issue
Stamatatos, E., & Koppel, M. (2011) · Language Resources and Evaluation · doi:10.1007/s10579-011-9136-1 · Cited by 45 · cited 1×
“…(Chen and Shu, 2024a,b, Hanley and Durumeric, 2024, Shu et al., 2020, Stiff and Johansson, 2022), protecting intellectual property rights (Meyer zu Eissen et al., 2007, Stamatatos and Koppel, 2011), and identifying fraudulent practices such as fake reviews (Afroz et al., 2012, Ott et al., 2011).”
CHEAT: A Large-scale Dataset for Detecting ChatGPT-writtEn AbsTracts
Yu, P., Chen, J., Feng, X., & Xia, Z. (2023) · arXiv · doi:10.48550/arxiv.2304.12008 · Cited by 45
Unsupervised and Distributional Detection of Machine-Generated TextPreprint
Gallé, M., Rozen, J., Kruszewski, G., & Elsahar, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.02878 · Cited by 42 · cited 1×
“Gallé et al. (2021) introduced an unsupervised method that identifies the over-appearance of repeated higher-order n-grams, distinguishing them from human-generated text.”
Lund, B. D., Wang, T., Mannuru, N. R., et al. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4389887 · Cited by 42
Cross-domain detection of GPT-2-generated technical text
Rodriguez, J., Hay, T., Gros, D., Shamsi, Z., & Srinivasan, R. (2022) · Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2022.naacl-main.88 · Cited by 40
Similarity Learning for Authorship Verification in Social Media
Boenninghoff, B., Nickel, R. M., Zeiler, S., & Kolossa, D. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683405 · Cited by 40 · cited 3×
“Moreover, syntax-augmented CNN models (Zhang et al., 2018), Convolutional Siamese Networks (Saedi and Dras, 2021), and attention-based Siamese Networks (Boenninghoff et al., 2019a) were explored.”
Text messaging forensics : Txt 4n6: idiolect free authorship analysis?
Grant, T. D. (2010) · doi:10.4324/9780203855607-50 · Cited by 40 · cited 1×
“In forensic investigations, authorship attribution plays a crucial role in solving murder cases disguised as suicides (Chaski, 2005, Grant, 2020), tracking terrorist threats (Cafiero and Camps, 2023, Winter, 2019), and aiding general criminal investigations (Argamon, 2018, Belvisi et al., 2020,…”
Sarvazyan, A. M., González, J. Á., Franco-Salvador, M., Rangel, F., Chulvi, B., & Rosso, P. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.11285 · Cited by 39
Smaller Language Models are Better Black-box Machine-Generated Text DetectorsPreprint
Mireshghallah, N., Mattern, J., Gao, S., Shokri, R., & Berg-Kirkpatrick, T. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.09859 · Cited by 39 · cited 1×
“Surrogate models, which are often small language models, are also applied to train classifiers (Mireshghallah et al., 2023, Verma et al., 2023).”
Token Prediction as Implicit Classification to Identify LLM-Generated Text
Chen, Y., Kang, H., Zhai, V., Li, L., Singh, R., & Raj, B. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.810 · Cited by 39 · cited 1×
“The primary goal is to determine whether a given piece of text was created by a specific human or by one of several LLMs (Beigi et al., 2024, Chen et al., 2023b, He et al., 2023, Soto et al., 2024, Uchendu et al., 2020, Venkatraman et al., 2023).”
Hanley, H. W. A., & Durumeric, Z. (2023) · arXiv · doi:10.48550/arxiv.2305.09820 · Cited by 36 · cited 2×
“…measures have been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law…”
The Importance of Suppressing Domain Style in Authorship AnalysisPreprint
Bischoff, S., Deckers, N., Schliebs, M., Thies, B., Hagen, M., Stamatatos, E., Stein, B., & Potthast, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.14714 · Cited by 36 · cited 1×
“• Fanfiction: Collected from fanfiction.net, this dataset includes fan-written fiction (Bischoff et al., 2020, Kestemont et al., 2021), providing insights into creative writing and authorship attribution in fictional narratives.”
Dudău, D. P., & Sava, F. A. (2021) · Frontiers in Psychology · doi:10.3389/fpsyg.2021.570568 · Cited by 36 · cited 1×
“…(LIWC) tool exemplifies how automated text analysis can use more than 100 psychological dimensions to analyze word use, reflecting distinct language variations among different groups in specific genres and languages (Dudău and Sava, 2021, Goldstein-Stewart et al., 2009, Pennebaker et al., 2015).”
Hanley, H. W. A., & Durumeric, Z. (2024) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v18i1.31333 · Cited by 36
Authorship Verification: A Review of Recent Advances
Stamatatos, E. (2016) · Research in Computing Science · doi:10.13053/rcs-123-1-1 · Cited by 34 · cited 1×
“Authorship verification determines whether a piece of writing was written by a specific individual (Stamatatos, 2016), while profiling infers characteristics such as age or gender from the author’s writing style (Argamon et al., 2009).”
ArguGPT: evaluating, understanding and identifying argumentative essays generated by GPT models
Liu, Y., Zhang, Z., Zhang, W., Yue, S., Zhao, X., Cheng, X., Zhang, Y., & Hu, H. (2023) · arXiv · doi:10.48550/arxiv.2304.07666 · Cited by 32
Provable Robust Watermarking for AI-Generated Text
Zhao, X., Ananth, P., Li, L., & Wang, Y.-X. (2023) · arXiv · doi:10.48550/arxiv.2306.17439 · Cited by 30 · cited 2×
“(Kirchenbauer et al., 2023, Meral et al., 2009, Topkara et al., 2005, Zhao et al., 2023).”
Can Large Language Models Identify Authorship?
Huang, B., Chen, C., & Shu, K. (2024) · arXiv · doi:10.48550/arxiv.2403.08213 · Cited by 30
Self-citationSiamese networks for large-scale author identification
Saedi, C., & Dras, M. (2021) · Computer Speech & Language · doi:10.1016/j.csl.2021.101241 · Cited by 30 · cited 1×
“Moreover, syntax-augmented CNN models (Zhang et al., 2018), Convolutional Siamese Networks (Saedi and Dras, 2021), and attention-based Siamese Networks (Boenninghoff et al., 2019a) were explored.”
DiPmark: A Stealthy, Efficient and Resilient Watermark for Large Language ModelsPreprint
Wu, Y.-H., Hu, Z.-M., Zhang, H., & Huang, H. (2023) · arXiv.org · doi:10.48550/arxiv.2310.07710 · Cited by 30
Identifying computer-generated text using statistical analysis
Nguyen-Son, H.-Q., Tieu, N.-D. T., Nguyen, H. H., Yamagishi, J., & Zen, I. E. (2017) · 2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipa.2017.8282270 · Cited by 30
MGTBench: Benchmarking Machine-Generated Text DetectionPreprint
He, X., Shen, X., Chen, Z., Michael Backes, A., & Zhang, Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2303.14822 · Cited by 29
Bao, G., Zhao, Y., Teng, Z., Yang, L., & Zhang, Y. (2023) · arXiv · doi:10.48550/arxiv.2310.05130 · Cited by 28 · cited 1×
“(Bao et al., 2023) enhanced this approach and proposed Fast-DetectGPT, which increases efficiency by using conditional probability curvature.”
Spear Phishing With Large Language Models
Hazell, J. (2023) · arXiv · doi:10.48550/arxiv.2305.06972 · Cited by 28 · cited 3×
“Future research should focus on refining existing authorship attri-bution methods to detect and prevent malicious activities such as generating misinformation, plagiarism, and propaganda (Goldstein et al., 2023, Hazell, 2023, Lund et al., 2023, Spitale et al., 2023).”
Breaking the Closed-World Assumption in Stylometric Authorship Attribution
Stolerman, A., Overdorf, R., Afroz, S., & Greenstadt, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-662-44952-3_13 · Cited by 28
Computer-Generated Text Detection Using Machine Learning: A Systematic Review
Beresneva, D. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-41754-7_43 · Cited by 27 · cited 1×
“This task involves recognizing the nuances of multi-source authorship with fine-grained precision.”
Improving author verification based on topic modeling
Potha, N., & Stamatatos, E. (2019) · Journal of the Association for Information Science and Technology · doi:10.1002/asi.24183 · Cited by 27 · cited 1×
“Key stylometric features include character and word frequencies (Sharma et al., 2018), parts of speech (Sundararajan and Woodard, 2018), punctuation, topics (Halvani and Graner, 2021, Potha and Stamatatos, 2019, Seroussi et al., 2014), and vocabulary richness.”
G3Detector: General GPT-Generated Text DetectorPreprint
Zhan, H., He, X., Xu, Q., Wu, Y., & Stenetorp, P. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.12680 · Cited by 27 · cited 1×
“(Zhan et al., 2023) developed G3Detector by fine-tuning RoBERTa-large for the same purpose.”
Learning Universal Authorship Representations
Rivera-Soto, R. A., Miano, O. E., Ordonez, J., Chen, B. Y., Khan, A., Bishop, M., & Andrews, N. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.70 · Cited by 25 · cited 7×
“These approaches, while offering higher performance, often sacrifice explainability for accuracy (Rivera-Soto et al., 2021).”
InfluentialSame Author or Just Same Topic? Towards Content-Independent Style Representations
Wegmann, A., Schraagen, M., & Nguyen, D. (2022) · arXiv · doi:10.48550/arxiv.2204.04907 · Cited by 25 · cited 1×
“Adaptation through style transfer has not resolved cross-domain issues (Boenninghoff et al., 2019b, Wegmann et al., 2022).”
Authorship Attribution of Social Media Messages
Theophilo, A., Giot, R., & Rocha, A. (2023) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2021.3123895 · Cited by 25
M4GT-Bench: Evaluation Benchmark for Black-Box Machine-Generated Text Detection
Wang, Y., Mansurov, J., Ivanov, P., Su, J., Shelmanov, A., Tsvigun, A., Mohammed Afzal, O., Mahmoud, T., Puccetti, G., Arnold, T., et al. (2024) · Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2024.acl-long.218 · Cited by 25
Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text
Hans, A., Schwarzschild, A., Cherepanova, V., Kazemi, H., Saha, A., Goldblum, M., Geiping, J., & Goldstein, T. (2024) · arXiv · doi:10.48550/arxiv.2401.12070 · Cited by 24 · cited 3×
“…measures have been employed for detection, including entropy (Gehrmann et al., 2019, Lavergne et al., 2008), perplexity (Beres-neva, 2016, Hans et al., 2024), average log-probability score (So-laiman et al., 2019), fluency (Holtzman et al., 2020), and Zipf’s word frequency law (Piantadosi,…”
Forensic Authorship Analysis of Microblogging Texts Using N-Grams and Stylometric Features
Belvisi, N. M. S., Muhammad, N., & Alonso‐Fernandez, F. (2020) · International Workshop on Biometrics and Forensics (IWBF) · doi:10.1109/iwbf49977.2020.9107953 · Cited by 24
On the Zero-Shot Generalization of Machine-Generated Text DetectorsPreprint
Pu, X., Zhang, J. (., Han, X.-C., et al. (2023) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2310.05165 · Cited by 24 · cited 2×
“LLM-generated text detectors often struggle to generalize to unseen domains encountered during training (Li et al., 2023a, Pu et al., 2023b, Rodriguez et al., 2022) and tend to perform better on LLMs they were specifically trained on (Chakraborty et al., 2023b, Li et al., 2023a, Pu et al., 2023b).”
DNA-GPT: Divergent N-Gram Analysis for Training-Free Detection of GPT-Generated Text
Yang, X., Cheng, W., Wu, Y., Petzold, L., Wang, W. Y., & Chen, H. (2023) · arXiv · doi:10.48550/arxiv.2305.17359 · Cited by 23 · cited 3×
“Additional methods leverage n-grams (Yang et al., 2023a), Uniform Information Density (UID) (Venkatraman et al., 2023), log rank information (Su et al., 2023b), and various linguistic features such as part-of-speech deter-miners, conjunctions, auxiliary relations, vocabulary, and emotional tone…”
Improved Certified Defenses against Data Poisoning with (Deterministic) Finite Aggregation
Wang, W., Levine, A., & Feizi, S. (2022) · arXiv · doi:10.48550/arxiv.2202.02628 · Cited by 22 · cited 1×
“Neural network-based detectors are vulnerable to adversarial and poisoning attacks (Goodfellow et al., 2015, Pu et al., 2023a, Wang et al., 2022).”
Person Identification from Text and Speech Genre Samples
Goldstein-Stewart, J., Winder, R., & Sabin, R. E. (2009) · Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics on - EACL '09 · doi:10.3115/1609067.1609104 · Cited by 22
On the Generalization of Training-based ChatGPT Detection MethodsPreprint
Xu, H., Ren, J., He, P.-F., et al. (2023) · Conference on Empirical Methods in Natural Language Processing · doi:10.48550/arxiv.2310.01307 · Cited by 22
Stylometric Authorship Attribution of Collaborative Documents
Dauber, E., Overdorf, R., & Greenstadt, R. (2017) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-60080-2_9 · Cited by 22 · cited 2×
“Collaborative texts, which blend multiple writing styles, further mask individual contributions and obscure distinct authorial signals (Dauber et al., 2017).”
GPT-Sentinel: Distinguishing Human and ChatGPT Generated Content
Chen, Y., Kang, H., Zhai, V., Li, L., Singh, R., & Raj, B. (2023) · arXiv · doi:10.48550/arxiv.2305.07969 · Cited by 21 · cited 2×
“Additionally, neural network-based models like BERT (Ippolito et al., 2020), Radar (Hu et al., 2023), and GPT-sentinel (Chen et al., 2023a) can also be applied.”
Large Language Models can be Guided to Evade AI-Generated Text Detection
Lu, N., Liu, S., He, R., Wang, Q., Ong, Y.-S., & Tang, K. (2023) · arXiv · doi:10.48550/arxiv.2305.10847 · Cited by 21 · cited 2×
“…also lack robustness to various factors, such as alternative decoding strategies (Ippolito et al., 2020), input sequence length (Solaiman et al., 2019), different prompts (Kumarage et al., 2023, Lu et al., 2023), repetition penalties (Fishchuk and Braun, 2023), and human edits (Zhang et al., 2024).”
GPT-who: An Information Density-based Machine-Generated Text Detector
Venkatraman, S., Uchendu, A., & Lee, D. (2023) · arXiv · doi:10.48550/arxiv.2310.06202 · Cited by 21 · cited 2×
“This simplifies the task to either LLM-generated Text Detection or LLM-generated Text Attribution (Crothers et al., 2023, Yang et al., 2023b).”
Learning Invariant Representations of Social Media Users
Andrews, N., & Bishop, M. (2019) · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) · doi:10.18653/v1/d19-1178 · Cited by 19 · cited 2×
“Additionally, combinations of convolu-tions and transformers have been employed to learn embeddings for comparison tasks (Andrews and Bishop, 2019).”
On pushing DeepFake Tweet Detection capabilities to the limits
Gambini, M., Fagni, T., Falchi, F., & Tesconi, M. (2022) · 14th ACM Web Science Conference 2022 · doi:10.1145/3501247.3531560 · Cited by 19 · cited 1×
“…LLM-generated text is crucial for several applications, including combating misinformation and disinformation on social media (Chen and Shu, 2024b, Gambini et al., 2022, Stiff and Johansson, 2022), identifying spam (Jindal and Liu, 2008), preventing phishing attacks (Hazell, 2023), identifying…”
M4: Multi-generator, Multi-domain, and Multi-lingual Black-Box Machine-Generated Text Detection
Wang, Y., Mansurov, J., Ivanov, P., Su, J., Shelmanov, A., Tsvigun, A., Whitehouse, C., Afzal, O. M., Mahmoud, T., Sasaki, T., et al. (2023) · arXiv · doi:10.48550/arxiv.2305.14902 · Cited by 19
LLM-as-a-Coauthor: Can Mixed Human-Written and Machine-Generated Text Be Detected?
Zhang, Q., Gao, C., Chen, D., Huang, Y., Huang, Y., Sun, Z., Zhang, S., Li, W., Fu, Z., Wan, Y., et al. (2024) · Findings of the Association for Computational Linguistics: NAACL 2024 · doi:10.18653/v1/2024.findings-naacl.29 · Cited by 18 · cited 10×
“Neural network-based detectors, generally outperform metric-based methods in both human authorship attribution and LLM-generated text detection problems (He et al., 2023, Zhang et al., 2024).”
InfluentialPOSNoise: An Effective Countermeasure Against Topic Biases in Authorship Analysis
Halvani, O., & Graner, L. (2020) · ARES · doi:10.1145/3465481.3470050 · Cited by 18 · cited 1×
“Key stylometric features include character and word frequencies (Sharma et al., 2018), parts of speech (Sundararajan and Woodard, 2018), punctuation, topics (Halvani and Graner, 2021, Potha and Stamatatos, 2019, Seroussi et al., 2014), and vocabulary richness.”
Ghostbuster: Detecting Text Ghostwritten by Large Language Models
Verma, V., Fleisig, E., Tomlin, N., & Klein, D. (2023) · arXiv · doi:10.48550/arxiv.2305.15047 · Cited by 16 · cited 2×
“There are fewer studies on this task due to its difficulty, and existing research often makes simplifications.”
How effectively can machines defend against machine-generated fake news? An empirical study
Bhat, M. M., & Parthasarathy, S. (2020) · Proceedings of the First Workshop on Insights from Negative Results in NLP · doi:10.18653/v1/2020.insights-1.7 · Cited by 16 · cited 2×
“These perturbations have diminished the effectiveness of existing text detection methods (Bhat and Parthasarathy, 2020).”
Paraphrase Detection: Human vs. Machine ContentPreprint
Becker, J., Wahle, J. P., Ruas, T., et al. (2023) · arXiv.org · doi:10.48550/arxiv.2303.13989 · Cited by 16 · cited 1×
“…including homoglyph attacks (Gagiano et al., 2021, Macko et al., 2024), whitespace insertion (Cai and Cui, 2023), syntactic perturbations (Bhat and Parthasarathy, 2020), synonym replacement (Kulkarni et al., 2023), and paraphrasing (Becker et al., 2023, Krishna et al., 2024, Shi et al., 2024).”
Liu, Z., Yao, Z., Li, F., & Luo, B. (2023) · arXiv · doi:10.48550/arxiv.2306.05524 · Cited by 16
Gehrmann, S., Strobelt, H., Rush, A. M., et al. (2019) · doi:10.18653/v1/p19-3 · Cited by 14
Mosca, E., Abdalla, M. H. I., Basso, P., Musumeci, M., & Groh, G. (2023) · Proceedings of the 3rd Workshop on Trustworthy Natural Language Processing (TrustNLP 2023) · doi:10.18653/v1/2023.trustnlp-1.17 · Cited by 14
A Survey on Data Selection for Language Models
Albalak, A., Elazar, Y., Xie, S. M., Longpre, S., Lambert, N., Wang, X., Muennighoff, N., Hou, B., Pan, L., Jeong, H., et al. (2024) · arXiv · doi:10.48550/arxiv.2402.16827 · Cited by 12 · cited 1×
“Moreover, leveraging LLMs for data augmentation and synthetic data generation can create diverse training datasets, which in turn improves the generalization of attribution models (Albalak et al., 2024).”
Who could be behind QAnon? Authorship attribution with supervised machine-learningPreprint
Cafiero, F., & Camps, J.-B. (2023) · Digital Scholarship in the Humanities · doi:10.48550/arxiv.2303.02078 · Cited by 12 · cited 1×
“…authorship attribution plays a crucial role in solving murder cases disguised as suicides (Chaski, 2005, Grant, 2020), tracking terrorist threats (Cafiero and Camps, 2023, Winter, 2019), and aiding general criminal investigations (Argamon, 2018, Belvisi et al., 2020, Koppel et al., 2008).”
The Topic Confusion Task: A Novel Evaluation Scenario for Authorship Attribution
Altakrori, M., Cheung, J. C. K., & Fung, B. C. M. (2021) · Findings of the Association for Computational Linguistics: EMNLP 2021 · doi:10.18653/v1/2021.findings-emnlp.359 · Cited by 11 · cited 1×
“Attribution performance tends to drop when known and query texts differ in topic or genre (Altakrori et al., 2021).”
Through the Looking Glass: Learning to Attribute Synthetic Text Generated by Language Models
Munir, S., Batool, B., Shafiq, Z., Srinivasan, P., & Zaffar, F. (2021) · Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume · doi:10.18653/v1/2021.eacl-main.155 · Cited by 11 · cited 2×
“Variations in model architecture (such as the number of layers and parameters), training methods (including pre-training and fine-tuning), and generation techniques (like sampling parameters) all subtly influence the characteristics of the generated texts (Munir et al., 2021).”
How Reliable Are AI-Generated-Text Detectors? An Assessment Framework Using Evasive Soft Prompts
Kumarage, T., Sheth, P., Moraffah, R., Garland, J., & Liu, H. (2023) · arXiv · doi:10.48550/arxiv.2310.05095 · Cited by 11 · cited 1×
“These texts often start as human drafts and are later extended or revised by LLMs, or the process might occur in reverse.”
Chakraborty, M., Tonmoy, S. M. T. I., Zaman, S. M. M., Sharma, K., Barman, N. R., Gupta, C., Gautam, S., Kumar, T., Jain, V., Chadha, A. R., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2310.05030 · Cited by 11 · cited 3×
“Analyzing and classifying texts co-authored by humans and LLMs presents a significant challenge due to the blending of human and machine stylistic features.”
Learning Interpretable Style Embeddings via Prompting LLMs
Patel, A., Rao, D., Kothary, A., McKeown, K., & Callison-Burch, C. (2023) · arXiv · doi:10.48550/arxiv.2305.12696 · Cited by 10 · cited 2×
“Traditionally, LLMs have been employed mainly for auxiliary tasks such as feature extraction and data annotation (Patel et al., 2023).”
MAGE: Machine-generated Text Detection in the Wild
Li, Y., Li, Q., Cui, L., Bi, W., Wang, Z., Wang, L., Yang, L., Shi, S., & Zhang, Y. (2023) · arXiv · doi:10.48550/arxiv.2305.13242 · Cited by 10
Detecting and Understanding Textual Deepfakes in Online Reviews
Kowalczyk, P., Röder, M., Dürr, A., & Thiesse, F. (2022) · Proceedings of the Annual Hawaii International Conference on System Sciences · doi:10.24251/hicss.2022.184 · Cited by 9 · cited 1×
“For instance, Kowalczyk et al. (2022) detected GPT-2-generated fake reviews using Shapley Additive Explanations (SHAP) (Lundberg and Lee, 2017).”
Who's Your Judge? On the Detectability of LLM-Generated Judgments
Li, D.-W., Tan, Z., Zhao, C., Jiang, B., Huang, B.-X., Ma, P., Alnaibari, A., Shu, K., & Liu, H. (2025) · arXiv.org · doi:10.48550/arxiv.2509.25154 · Cited by 9 · cited 1×
“…2022), identifying spam (Jindal and Liu, 2008), preventing phishing attacks (Hazell, 2023), identifying fake reviews (Salminen et al., 2022), detecting machine-generated scientific papers (Liu et al., 2023a, Rodriguez et al., 2022), and making academic peer review judgments (Li et al., 2025).”
Self-citationOrigin Tracing and Detecting of LLMs
Li, L., Wang, P., Ren, K., Sun, T., & Qiu, X. (2023) · arXiv · doi:10.48550/arxiv.2304.14072 · Cited by 8 · cited 2×
“A sub-problem is to attribute the text solely to LLMs, also known as model sourcing (Yang et al., 2023a) or origin tracing (Li et al., 2023b).”
HC3 Plus: A Semantic-Invariant Human ChatGPT Comparison CorpusPreprint
Su, Z., Wu, X. C., Zhou, W., Ma, G., & Hu, S. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.02731 · Cited by 8
Linking User Accounts across Social Media Platforms
Sinnott, R., & Wang, Z. (2021) · IEEE/ACM International Conference on Big Data Computing, Applications and Technologies (BDCAT) · doi:10.1145/3492324.3494157 · Cited by 8 · cited 1×
“…digital realm, authorship attribution helps safe-guard the integrity of content by preventing deceptive social media activities (Hazell, 2023), detecting account compromises (Barbon et al., 2017), and linking user profiles across various social networks (Shu et al., 2017, Sinnott and Wang, 2021).”
Bevendorff, J., Chulvi, B., Fersini, E., Heini, A., Kestemont, M., Kredens, K., Mayerl, M., Ortega-Bueno, R., Pęzik, P., Potthast, M., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-99739-7_42 · Cited by 8
Evaluating the Social Impact of Generative AI Systems
Solaiman, I., Talat, Z., Agnew, W. S., Ahmad, L., Baker, D., Blodgett, S. L., Chen, C., Daumé, H., Dodge, J., Duan, I., et al. (2025) · Oxford University Press eBooks · doi:10.1093/oxfordhb/9780198940272.013.0025 · Cited by 7 · cited 1×
“This inability to distinguish between human and machine-generated content un-dermines the integrity of authorship, complicates legal and ethical responsibilities, and threatens the credibility of digital content and the safety of online space (Solaiman et al., 2023, Vidgen et al., 2024).”
Self-citationEvade ChatGPT Detectors via A Single Space
Cai, S., & Cui, W. (2023) · arXiv · doi:10.48550/arxiv.2307.02599 · Cited by 7 · cited 1×
“…detectors are vulnerable to adversarial attacks, including homoglyph attacks (Gagiano et al., 2021, Macko et al., 2024), whitespace insertion (Cai and Cui, 2023), syntactic perturbations (Bhat and Parthasarathy, 2020), synonym replacement (Kulkarni et al., 2023), and paraphrasing (Becker et…”
Sharma, A., Nandan, A., & Ralhan, R. (2018) · arXiv · doi:10.48550/arxiv.1812.10281 · Cited by 7 · cited 1×
“Key stylometric features include character and word frequencies (Sharma et al., 2018), parts of speech (Sundararajan and Woodard, 2018), punctuation, topics (Halvani and Graner, 2021, Potha and Stamatatos, 2019, Seroussi et al., 2014), and vocabulary richness.”
Unified and Multilingual Author Profiling for Detecting Haters
Schlicht, I. B., & de Paula, A. F. M. (2021) · RiuNet (Universitat Politècnica de València) · doi:10.48550/arxiv.2109.09233 · Cited by 7 · cited 1×
“…architectures (Devlin et al., 2019) such as BERT (Fabien et al., 2020, Ippolito et al., 2020, Manolache et al., 2021), Sentence-BERT (Rivera-Soto et al., 2021, Schlicht and de Paula, 2021), and RoBERTa (Huertas-Tato et al., 2022), have proven effective for learning authorship representation.”
HANSEN: Human and AI Spoken Text Benchmark for Authorship Analysis
Tripto, N. I., Uchendu, A., Le, T., Setzu, M., Giannotti, F., & Lee, D. (2023) · arXiv · doi:10.48550/arxiv.2310.16746 · Cited by 6 · cited 1×
“Open-source detectors provide more transparent baselines, but they similarly face robustness and generalization challenges.”
Authorship Obfuscation in Multilingual Machine-Generated Text Detection
Macko, D., Moro, R., Uchendu, A., Srba, I., Lucas, J. S., Yamashita, M., Tripto, N. I., Lee, D., Simko, J., & Bielikova, M. (2024) · arXiv · doi:10.48550/arxiv.2401.07867 · Cited by 6
Introducing v0.5 of the AI Safety Benchmark from MLCommons
Vidgen, B., Agrawal, A., Ahmed, A. M., Akinwande, V., Al-Nuaimi, N., Alfaraj, N., Alhajjar, E., Aroyo, L., Bavalatti, T., Bartolo, M., et al. (2024) · arXiv · doi:10.48550/arxiv.2404.12241 · Cited by 5 · cited 1×
“This inability to distinguish between human and machine-generated content un-dermines the integrity of authorship, complicates legal and ethical responsibilities, and threatens the credibility of digital content and the safety of online space (Solaiman et al., 2023, Vidgen et al., 2024).”
Self-citationLong-form analogies generated by chatGPT lack human-like psycholinguistic properties
Seals, S. M., & Shalin, V. L. (2023) · arXiv · doi:10.48550/arxiv.2306.04537 · Cited by 5 · cited 1×
“…distinctively from lexical content words, revealing correlations with socio-cultural categories such as gender, age, and native language (Chambers et al., 2013, Nerbonne, 2014, Seals and Shalin, 2023), which are pivotal in understanding identity and social affiliations (Argamon et al., 2009).”
PART: Pre-trained Authorship Representation Transformer
Huertas-Tato, J., Martin, A., & Camacho, D. (2022) · arXiv · doi:10.48550/arxiv.2209.15373 · Cited by 4
Rethinking the Authorship Verification Experimental SetupsPreprint
Brad, F., Manolache, A., Burceanu, E., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2112.05125 · Cited by 4
Efficient Black-Box Adversarial Attacks on Neural Text Detectors
Fishchuk, V., & Braun, D. (2023) · arXiv · doi:10.48550/arxiv.2311.01873 · Cited by 3 · cited 1×
“…also lack robustness to various factors, such as alternative decoding strategies (Ippolito et al., 2020), input sequence length (Solaiman et al., 2019), different prompts (Kumarage et al., 2023, Lu et al., 2023), repetition penalties (Fishchuk and Braun, 2023), and human edits (Zhang et al., 2024).”
Research works printed but not linked: 20
- [7]7. Shlomo Argamon. 2018. Computational forensic author- ship analysis: Promises and pitfalls. Language and Law/Linguagem e Direito 5, 2 (2018), 7-37.
- [16]16. Georgios Barlas and Efstathios Stamatatos. 2020b. Cross- domain authorship attribution using pre-trained language models. In Artificial Intelligence Applications and Innova- tions: 16th IFIP WG 12.5 International Conference, AIAI 2020, Neos Marmaras, Greece, June 5-7, 2020, Proceed- ings, Part I 16. Springer, 255-266.
- [54]54. Maël Fabien, Esau Villatoro-Tello, Petr Motlicek, and Shan- tipriya Parida. 2020. BertAA : BERT fine-tuning for Authorship Attribution. In Proceedings of the 17th In- ternational Conference on Natural Language Processing (ICON). NLP Association of India (NLPAI), Indian Insti- tute of Technology Patna, Patna, India, 127-137. https: //aclanthology.org/2020.icon-main.16
- [58]58. Rinaldo Gagiano, Maria Myung-Hee Kim, Xiuzhen Zhang, and Jennifer Biggs. 2021. Robustness Analysis of Grover for Machine-Generated News Detection. In Proceedings of the The 19th Annual Workshop of the Australasian Language Technology Association. Australasian Language Technology Association, Online, 119-127. https://aclanthology. org/2021.alta-1.12
- [90]90. Ganesh Jawahar, Muhammad Abdul-Mageed, and Laks Lakshmanan, V.S. 2020b. Automatic Detection of Ma- chine Generated Text: A Critical Survey. In Proceed- ings of the 28th International Conference on Computa- tional Linguistics. International Committee on Computa- tional Linguistics, Barcelona, Spain (Online), 2296-2309. https://doi.org/10.18653/v1/2020.coling-main.208 · DOI · Paper page
- [94]94. Prannay Khosla, Piotr Teterwak, Chen Wang, Aaron Sarna, Yonglong Tian, Phillip Isola, Aaron Maschinot, Ce Liu, and Dilip Krishnan. 2020. Supervised Contrastive Learn- ing. In Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6-12, 2020, virtual, Hugo Larochelle, Marc'Aurelio Ranzato, Raia Hadsell, Maria-Florina Balcan, and Hsuan-Tien Lin (Eds.). https://proceedings.neurips.cc/paper/2020/hash/ d89a66c7c80a29b1bdbab0f2a1a94af8-Abstract.html
- [99]99. Moshe Koppel, Jonathan Schler, and Eran Messeri. 2008. Au- thorship attribution in law enforcement scenarios. NATO Security Through Science Series D-Information and Com- munication Security 15 (2008), 111.
- [108]108. Thomas Lavergne, Tanguy Urvoy, and François Yvon. 2008. Detecting Fake Content with Relative Entropy Scoring. Pan 8, 27-31 (2008), 4.
- [123]123. David Madigan, Alexander Genkin, David D Lewis, Shlomo Argamon, Dmitriy Fradkin, and Li Ye. 2005b. Author identification on the large scale. In Proceedings of the 2005 Meeting of the Classification Society of North America (CSNA).
- [127]127. Rohith Menon and Yejin Choi. 2011. Domain Independent Authorship Attribution without Domain Adaptation. In Proceedings of the International Conference Recent Ad- vances in Natural Language Processing 2011. Association for Computational Linguistics, Hissar, Bulgaria, 309-315. https://aclanthology.org/R11-1043
- [145]145. Jiameng Pu, Zain Sarwar, Sifat Muhammad Abdullah, Ab- dullah Rehman, Yoonjin Kim, Parantapa Bhattacharya, Mobin Javed, and Bimal Viswanath. 2023a. Deepfake text detection: Limitations and opportunities. In 2023 IEEE Symposium on Security and Privacy (SP). IEEE, 1613-1630.
- [150]150. Vinay Venkatesh Ramasesh, Aitor Lewkowycz, and Ethan Dyer. 2022. Effect of scale on catastrophic forgetting in neural networks. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022. OpenReview.net. https://openreview. net/forum?id=GhVS8_yPeEa
- [155]155. Juan Rodriguez, Todd Hay, David Gros, Zain Shamsi, and Ravi Srinivasan. 2022b. Cross-Domain Detection of GPT-2-Generated Technical Text. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, Seattle, United States, 1213-1233. https: //doi.org/10.18653/v1/2022.naacl-main.88 · DOI · Paper page
- [163]163. Jonathan Schler, Moshe Koppel, Shlomo Argamon, and James W Pennebaker. 2006. Effects of age and gender on blogging.. In AAAI spring symposium: Computational approaches to analyzing weblogs, Vol. 6. 199-205.
- [168]168. Seroussi, Russell Smyth, and Ingrid Zukerman. 2011. Ghosts from the high court's past: Evidence from compu- tational linguistics for Dixon ghosting for Mctiernan and rich. University of New South Wales Law Journal, The 34, 3 (2011), 984-1005.
- [174]174. Prasha Shrestha, Sebastian Sierra, Fabio González, Manuel Montes, Paolo Rosso, and Thamar Solorio. 2017b. Con- volutional Neural Networks for Authorship Attribution of Short Texts. In Proceedings of the 15th Conference of the European Chapter of the Association for Computa- tional Linguistics: Volume 2, Short Papers. Association for Computational Linguistics, Valencia, Spain, 669-674. https://aclanthology.org/E17-2106
- [190]190. Jinyan Su, Terry Yue Zhuo, Di Wang, and Preslav Nakov. 2023c. Detectllm: Leveraging log rank information for zero- shot detection of machine-generated text. ArXiv preprint abs/2306.05540 (2023). https://arxiv.org/abs/2306. 05540 · arXiv
- [192]192. Kalaivani Sundararajan and Damon Woodard. 2018. What represents "style" in authorship attribution?. In Proceed- ings of the 27th International Conference on Computa- tional Linguistics. Association for Computational Linguis- tics, Santa Fe, New Mexico, USA, 2814-2822. https: //aclanthology.org/C18-1238
- [194]194. Ruixiang Tang, Yu-Neng Chuang, and Xia Hu. 2023b. The science of detecting llm-generated texts. ArXiv preprint abs/2303.07205 (2023). https://arxiv.org/abs/2303. 07205 · arXiv
- [216]216. Xianjun Yang, Wei Cheng, Yue Wu, Linda Petzold, William Yang Wang, and Haifeng Chen. 2023b. Dna-gpt: Divergent n-gram analysis for training-free detection of gpt-generated text. ArXiv preprint abs/2305.17359 (2023). https://arxiv.org/abs/2305.17359 · arXiv
Cited by
Sorted by citations
Artificial Intelligence and the Psychology of Human Connection
Boyd, R. L., & Markowitz, D. M. (2026) · Perspectives on Psychological Science · doi:10.1177/17456916251404394 · Cited by 78
“Prior work has shown that people often misattribute authorship— mistaking AI-generated writing for human writing, or vice versa—on the basis of subtle stylistic cues (Gunser et al., 2021; B. Huang et al., 2025; Köbis & Mossink, 2021).”
Advancements in AI-Generated Content Forensics: A Systematic Literature Review
Xu, Q., Mu, W., Li, J., Sun, T., & Jiang, X. (2025) · ACM Computing Surveys · doi:10.1145/3760526 · Cited by 19
“Beyond binary detection, a more profound challenge is model attribution–accurately identifying the source model of a piece of content [260].”
Auditing the Shadows: A Review of Methods to Detect Shared Training Data in Large Language Models
Naser, M. Z. (2025) · ACM Computing Surveys · doi:10.1145/3774897 · Cited by 7
“Each methodology is distinguished by its underlying assumptions about how training data manifests in learned parameters and by the computational strategies they employ to isolate signals that reliably indicate the presence of shared textual fragments [ 68 –70 ].”
Distinguishing AI-Generated and Human-Written Text Through Psycholinguistic Analysis
Opara, C. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-98462-4_27 · Cited by 3
Authorship Drift: How Self-Efficacy and Trust Evolve During LLM-Assisted Writing
Park, Y. S., Arvi, N. A. P., Kim, S., & Kim, J. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790276 · Cited by 8
Metadata
- Authors
- Huang, B., Chen, C., & Shu, K.
- Affiliations
- Illinois Institute of Technology (US)
- Emory University (US)
- Venue
- ACM SIGKDD Explorations Newsletter
- Published
- 2025-01-21
- Volume
- 26
- Issue
- 2
- Pages
- 21-43
- ISSN
- 1931-0145 (ISSN-L)
- 1931-0153
- PMC
- PMC12019761
- OpenAlex
- W4406939734
- Cited by
- 42
- References
- 209
- Citations by year
- 2026: 28 · 2025: 14
- FWCI
- 16.71
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- ACM SIGKDD Explorations Newsletter (publishedVersion)
- PubMed (publishedVersion)
- PubMed Central (acceptedVersion)
- Europe PMC (PubMed Central) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation (2346158, 2241068, 2339198)
- U.S. Department of Homeland Security (17STQAC00001-07-04)
- Cisco Systems
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Steganography and Watermarking Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Keywords
- authorship attribution, large language models, AI-generated text detection, misinformation, plagiarism detection, forensic investigation, explainability, domain generalization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
