The DOI matches its Crossref record; 21 of the 44 references it cites were found in a registry or an index.
Attribute or Abstain: Large Language Models as Long Document Assistants
Open accessAbstract
LLMs can help humans working with long documents, but are known to hallucinate.Attribution can increase trust in LLM responses: The LLM provides evidence that supports its response, which enhances verifiability.Existing approaches to attribution have only been evaluated in RAG settings, where the initial retrieval confounds LLM performance.This is crucially different from the long document setting, where retrieval is not needed, but could help.Thus, a long document specific evaluation of attribution is missing.To fill this gap, we present LAB, a benchmark of 6 diverse long document tasks with attribution, and experiments with different approaches to attribution on 5 LLMs of different sizes.We find that citation, i.e. response generation and evidence extraction in one step, performs best for large and fine-tuned models, while additional retrieval can help for small, prompted models.We investigate whether the "Lost in the Middle" phenomenon exists for attribution, but do not find this.We also find that evidence quality can predict response quality on datasets with simple responses, but not so for complex responses, as models struggle with providing evidence for complex claims.We release code and data for further investigation 1 .
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Moreover, the citations are generated conditioned on the response, which has been shown to be more effective compared to prior or post-hoc citations (Buchmann et al., 2024).”
From In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents (2025) · cites this paper as Buchmann et al., 2024
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“This finding is also consistent with related work such as Buchmann et al. (2024).”
From PeerQA: A Scientific Question Answering Dataset from Peer Reviews (2025) · cites this paper as Buchmann et al. (2024)
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 12
- Ran El-Yaniv and Yair Wiener. 2010. On the founda- tions of noise-free selective classification. Journal of Machine Learning Research, 11(53):1605-1641.
- Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. Lora: Low-rank adaptation of large language models. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022. OpenReview.net.
- Chin-Yew Lin. 2004. ROUGE: A package for auto- matic evaluation of summaries. In Text Summariza- tion Branches Out, pages 74-81, Barcelona, Spain. Association for Computational Linguistics.
- Shayne Longpre, Le Hou, Tu Vu, Albert Webson, Hyung Won Chung, Yi Tay, Denny Zhou, Quoc V Le, Barret Zoph, Jason Wei, and Adam Roberts. 2023. The flan collection: Designing data and methods for effective instruction tuning. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 22631-22648.
- PMLR.
- Ilya Loshchilov and Frank Hutter. 2019. Decoupled weight decay regularization. In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019. OpenRe- view.net.
- Jianmo Ni, Chen Qu, Jing Lu, Zhuyun Dai, Gustavo Hernandez Abrego, Ji Ma, Vincent Zhao, Yi Luan, Keith Hall, Ming-Wei Chang, and Yinfei Yang. 2022. Large dual encoders are generalizable retrievers. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9844-9855, Abu Dhabi, United Arab Emirates. As- sociation for Computational Linguistics.
- Nandan Thakur, Nils Reimers, Andreas Rücklé, Ab- hishek Srivastava, and Iryna Gurevych. 2021. BEIR: A heterogeneous benchmark for zero-shot evaluation of information retrieval models. In Thirty-fifth Con- ference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2).
- Yunshu Wu, Hayate Iso, Pouya Pezeshkpour, Nikita Bhutani, and Estevam Hruschka. 2024. Less is more for long document summary evaluation by LLMs. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Lin- guistics (Volume 2: Short Papers), pages 330-343, St. Julian's, Malta. Association for Computational Linguistics.
- Jingqing Zhang, Yao Zhao, Mohammad Saleh, and Pe- ter J. Liu. 2020a. PEGASUS: pre-training with ex- tracted gap-sentences for abstractive summarization. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, volume 119 of Proceedings of Machine Learning Research, pages 11328-11339.
- PMLR.
- Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020b. Bertscore: Eval- uating text generation with BERT. In 8th Inter- national Conference on Learning Representations,
Cited by
Sorted by citations
Song, M., Sim, S. H., Bhardwaj, R., Chieu, H. L., Majumder, N., & Poria, S. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2409.11242 · Cited by 22
In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents
Tan, Z., Yan, J., Hsu, I.-H., Han, R., Wang, Z., Le, L., Song, Y., Chen, Y., Palangi, H., Lee, G., et al. (2025) · Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2025.acl-long.413 · Cited by 16
On Mechanistic Circuits for Extractive Question-AnsweringPreprint
Basu, S., Morariu, V. I., Wang, Z., Rossi, R. A., Zhao, C., Feizi, S., & Manjunatha, V. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2502.08059 · Cited by 3
Ojha, A. (2026) · American Journal of Financial Technology and Innovation · doi:10.54536/ajfti.v4i1.8551 · Cited by 2
In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents
Tan, Z., Yan, J., Hsu, I.-H., Han, R., Wang, Z., Le, L. T., Song, Y., Chen, Y., Palangi, H., Lee, G., et al. (2025) · arXiv · Cited by 54
Metadata
- Authors
- Buchmann, J., Liu, X., & Gurevych, I.
- Affiliations
- Technische Universität Darmstadt (DE)
- Venue
- Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
- Published
- 2024
- Pages
- 8113-8140
- OpenAlex
- W4404782773
- Cited by
- 3
- References
- 0
- Citations by year
- 2026: 3
- FWCI
- 0.73
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- dx.doi.org
- Licence
- cc-by
- Locations
- Conference on Empirical Methods in Natural Language Processing (EMNLP) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion, cc-by)
- Technischen Universität Darmstadt (submittedVersion, cc-by)
- Technischen Universität Darmstadt (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Technischen Universität Darmstadt (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Keywords
- large language models, attribution, long documents, hallucinations, evidence extraction, citation, retrieval-augmented generation, quality of evidence, response quality, LLM benchmark
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.
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
