Interpreting Language Models with Contrastive Explanations
Open accessThe DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.
Abstract
Model interpretability methods are often used to explain NLP model decisions on tasks such as text classification, where the output space is relatively small.However, when applied to language generation, where the output space often consists of tens of thousands of tokens, these methods are unable to provide informative explanations.Language models must consider various features to predict a token, such as its part of speech, number, tense, or semantics.Existing explanation methods conflate evidence for all these features into a single explanation, which is less interpretable for human understanding.To disentangle the different decisions in language modeling, we focus on explaining language models contrastively: we look for salient input tokens that explain why the model predicted one token instead of another.We demonstrate that contrastive explanations are quantifiably better than non-contrastive explanations in verifying major grammatical phenomena, and that they significantly improve contrastive model simulatability for human observers.We also identify groups of contrastive decisions where the model uses similar evidence, and we are able to characterize what input tokens models use during various language generation decisions.1
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
Lipton, Z. C. (2018) · Queue · doi:10.1145/3236386.3241340 · Cited by 5,427
Towards A Rigorous Science of Interpretable Machine Learning
Doshi-Velez, F., & Kim, B. (2017) · arXiv · doi:10.48550/arxiv.1702.08608 · Cited by 3,121
Stepin, I., Alonso, J. M., Catala, A., & Pereira-Farina, M. (2021) · IEEE Access · doi:10.1109/access.2021.3051315 · Cited by 376
GPT-Neo: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow
Black, S., Leo, G., Wang, P., Leahy, C., & Biderman, S. (2021) · Zenodo · doi:10.5281/zenodo.5297715 · Cited by 324
Evaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?
Hase, P., & Bansal, M. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.491 · Cited by 254
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
Contrastive Explanations for Model Interpretability
Jacovi, A., Swayamdipta, S., Ravfogel, S., Elazar, Y., Choi, Y., & Goldberg, Y. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.120 · Cited by 83
Contrastive and counterfactual reasoning in causal judgment.
McGill, A. L., & Klein, J. G. (1993) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.64.6.897 · Cited by 56
(2021) · doi:10.18653/v1/2021.conll-1.15 · Cited by 51
Cited by
Faithful Chain-of-Thought Reasoning
Lyu, Q., Havaldar, S., Stein, A., Zhang, L., Rao, D., Wong, E., Apidianaki, M., & Callison-Burch, C. (2023) · Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2023.ijcnlp-main.20 · Cited by 201
Dissecting Recall of Factual Associations in Auto-Regressive Language Models
Geva, M., Bastings, J., Filippova, K., & Globerson, A. (2023) · Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2023.emnlp-main.751 · Cited by 53
Dissecting Recall of Factual Associations in Auto-Regressive Language Models
Geva, M., Bastings, J., Filippova, K., & Globerson, A. (2023) · arXiv · doi:10.48550/arxiv.2304.14767 · Cited by 53
Facilitating Human-LLM Collaboration through Factuality Scores and Source Attributions
Do, H. J., Ostrand, R., Weisz, J. D., Dugan, C., Sattigeri, P., Wei, D., Murugesan, K., & Geyer, W. (2024) · arXiv · doi:10.48550/arxiv.2405.20434 · Cited by 12
Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations
Chen, Y., Zhong, R., Ri, N., Zhao, C., He, H., Steinhardt, J., Yu, Z., & McKeown, K. (2023) · arXiv · doi:10.48550/arxiv.2307.08678 · Cited by 5
Metadata
- Authors
- Yin, K., & Neubig, G.
- Venue
- Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
- Published
- 2022
- Pages
- 184-198
- OpenAlex
- W4385572952
- Cited by
- 37
- References
- 9
- FWCI
- 5.24
- Type
- conference paper
- Language
- en
- Open access
- hybrid
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Interpreting Language Models with Contrastive Explanations. https://app.cochat.ai/papers/10.18653/v1/2022.emnlp-main.14
- The paper
- Yin, K., & Neubig, G. (2022). Interpreting Language Models with Contrastive Explanations. Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. https://doi.org/10.18653/v1/2022.emnlp-main.14
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.