Document Attribution: Examining Citation Relationships using Large Language Models

Franck Dernoncourt; Nedim Lipka; Ryan A. Rossi; Vipula Rawte

REVIEW 4 major objections 6 minor 14 references

Document Attribution: Examining Citation Relationships using Large Language Models

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Framing attribution as a textual entailment question lets a zero-shot LLM match or beat prior attribution benchmarks without fine-tuning.

desk verdict Simple entailment prompt for attribution is worth a quick test, but the paper's own tables don't support the claimed improvements, and the attention analysis reduces to an always-positive classifier. read the letter →

arxiv 2505.06324 v1 pith:LJ73C25Z submitted 2025-05-09 cs.IR cs.AI

Vipula Rawte , Ryan A. Rossi , Franck Dernoncourt , Nedim Lipka This is my paper · ORCID

classification cs.IRcs.AI

keywords attributiontextualentailmentzero-shotpromptinglargelanguagemodelscitationverificationBenchFLAN-UL2attentionlayers

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that LLM attribution—deciding whether a generated claim is supported by a cited reference—can be handled by a simple zero-shot textual entailment prompt. Using the prompt 'Does the REFERENCE entail the CLAIM?' with the FLAN-UL2 model, it reports the best average $F_1$ on both the in-distribution (73.8) and out-of-distribution (83.43) sets of AttributionBench, edging out the strongest published baselines (73.6 and 81.3). The paper also reports a preliminary attention-based classifier using flan-t5-small on one subset, where attention features from most layers beat the zero-shot baseline $F_1$. The intended upshot is that expensive fine-tuning and specialized training are not necessary for basic citation-claim verification, and that attention layers carry usable attribution signal.

What carries the argument

The central object is the textual entailment relation $\mathrm{Entail}(S_1, S_2)$, implemented as a zero-shot prompt: 'Answer the question with ONLY a YES or NO. Does the REFERENCE entail the CLAIM?' The reference text is the premise, the claim is the hypothesis, and the model's yes/no decision is the attribution label. This collapses attribution into a judgment the LLM can make from its pretrained language understanding, without task-specific fine-tuning. For the attention analysis, the machinery is a per-layer binary classifier: attention weights from each of flan-t5-small's 12 layers feed a fully connected layer that predicts the attribution class, isolating where attribution signal lives.

What would settle it

Rerun the AttributionBench zero-shot FLAN-UL2 baseline and the paper's entailment prompt under the same decoding settings and metric implementation; if the baseline reproduces at 73.8 or above, or if repeated runs of the entailment prompt vary by more than 0.2 F1 points, the claimed in-distribution improvement is not demonstrated.

Watch

Extended reading notes

Core claim

The paper's central claim is that the binary attribution task defined in AttributionBench—whether a claim is attributable to a given reference—is adequately solved by reframing it as textual entailment. With the reference as premise and the claim as hypothesis, an instruction-tuned LLM asked a yes/no entailment question yields $F_1 = 73.8$ on the in-distribution average and $F_1 = 83.43$ on the out-of-distribution average, compared with the best baseline values of 73.6 and 81.3. On individual subsets, the method sets the best reported scores on Stanford-GenSearch and LFQA (ID) and on AttrScore-GenSearch and HAGRID (OOD). For the attention half, the paper claims that with flan-t5-small on the LFQA subset, linear classifiers over layer-wise attention weights beat the zero-shot baseline $F_1$ at most layers, with exceptions at layer 4 and layers 8 through 11.

Load-bearing premise

The load-bearing premise is that the published baseline F1 values were computed under exactly the same evaluation conditions as the paper's own runs, so the reported gains—especially the 0.2-point in-distribution difference—reflect the method rather than protocol or sampling noise.

Editorial extensions

If this is right

Attribution can be treated as a textual entailment problem, so any instruction-tuned LLM capable of yes/no entailment judgments can be used for citation verification without training data.
The reported out-of-distribution average (83.43) suggests the entailment formulation transfers across datasets with different claim-and-reference styles better than the previous zero-shot baselines.
Because the attention classifier performs above baseline on most layers of flan-t5-small, attribution signal appears distributed across the model rather than concentrated in a single layer, which could inform where to probe or intervene.
Simple prompt templates are enough to improve on established benchmarks; the gains do not require specialized attribution modules or retrieval changes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The in-distribution gain (0.2 $F_1$ points) is smaller than typical run-to-run variation on such benchmarks, so the stronger out-of-distribution gain (2.1 points) is the more informative result, but it rests on comparing with published numbers rather than a re-run baseline.
The layer-wise attention results should be read cautiously: they come from one 168-instance subset with balanced classes but extreme false-positive and false-negative values at several layers, so they show feasibility, not a stable mechanism.
A natural testable extension would be to use the same entailment prompt on a per-sentence citation task and check whether yes/no judgments align with human citation-quality judgments, or to combine attention features from multiple layers to see whether the dip at layers 8 through 11 disappears.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here's my read on 2505.06324. The idea is clean: frame attribution as textual entailment with the prompt "Does the REFERENCE entail the CLAIM?" and test it on AttributionBench. That is a legitimate, cheap thing to measure, and the authors are explicit that they follow the same zero-shot setup as the benchmark, which makes cross-paper comparison meaningful in principle. The related work is on point, and the paper is easy to read.

The problems are in the numbers. Table 1 lists flan-ul2 per-dataset F1s of 55, 75.2, 84.16, and 85.38, but the reported ID average is 73.8. The simple mean is 74.9. That's an internal inconsistency in the headline result. Fixing it makes the claimed ID gain bigger (about 1.3 points over the published baseline), but there are still no error bars, seeds, or significance tests, so the gain is unreplicated and within plausible noise. The abstract's 2.4% OOD figure is also mis-sourced: it matches a comparison to the authors' own gpt4-o row (81.48) rather than the best published baseline (81.3). That's a minor issue, but it matters for honesty of reporting.

The attention section is the weakest part. On the balanced 84/84 LFQA subset, any classifier that predicts "yes" for everything gets F1=66.67. Many rows in Table 3 have FP=100, FN=0, which is exactly that degenerate solution. The paper presents these as "outperforming the baseline" when the zero-shot F1 is only 20, but 66.67 is not evidence that attention helps; it is evidence that the probe collapsed. No train/test split is described for the per-layer linear classifiers, so overfitting is a real risk. The Limitations section concedes compute constraints, which explains the small scale but not the arithmetic error or the missing experimental controls.

So: the prompt variant is worth a quick replication by anyone already running AttributionBench, but as a paper the central claim is not established. I would not cite it as evidence without running my own control first. The right editorial decision is to desk reject in the current form, while inviting a resubmission that fixes the table arithmetic, reports variance, and replaces the attention analysis with a non-degenerate probe (majority baseline, held-out split). The underlying question is real; the evidence here is not.

Referee Report

4 major / 6 minor

Summary. This paper addresses automatic attribution evaluation, formulated as the binary task of deciding whether a claim is supported by a reference document, as instantiated by the AttributionBench benchmark. The authors propose two techniques. First, a zero-shot prompting method that recasts attribution as textual entailment ("Does the REFERENCE entail the CLAIM?"), evaluated with gpt4-o and flan-ul2 on the ID and OOD subsets of AttributionBench; the abstract claims a 0.27% ID and a 2.4% OOD relative improvement over the best baselines. Second, an attention-based analysis in which the attention weights of each layer of flan-t5-small are fed to a fully connected binary classifier, with per-layer F1 reported on the LFQA subset. The paper concludes that the entailment prompt outperforms existing zero-shot baselines and that attention layers, except layer 4 and layers 8 through 11, improve attribution.

Significance. If the reported results were statistically robust, the paper would offer a genuinely simple and reproducible zero-shot baseline for attribution: the prompt template is minimal, the evaluation uses the fixed public AttributionBench labels, and the method would be cheap to deploy. The attention analysis, however, is explicitly preliminary. The main limitation on significance is the size of the reported effects: a 0.2-point ID F1 gain on roughly 1,610 examples and OOD gains on subsets of 162 to 1,013 examples are within the range of typical evaluation noise, and all numbers come from single unseeded runs. Section 6 is candid about computational constraints, but the claims in the abstract outrun the statistical evidence in Sections 4.1 and 4.2. The contributions are potentially useful as an empirical note, not as an established improvement over the state of the art.

major comments (4)

[Table 1 and §4.2] The reported ID average for the proposed flan-ul2 row is arithmetically inconsistent with the row's own per-dataset F1 values: (55 + 75.2 + 84.16 + 85.38) / 4 = 74.9, not the reported 73.8. Because the abstract's headline "0.27% improvement" is the relative difference between 73.8 and the FLAN-UL2 baseline of 73.6, the central ID claim rests on an internally inconsistent number. The average must be recomputed consistently, and either the per-dataset values or the reported average corrected.
[Abstract, Table 2, §4.2] The claimed 2.4% OOD improvement is not stated against the best baseline. The best published OOD zero-shot baseline in Table 2 is FLAN-UL2 at 81.3, against which 83.43 is a 2.62% relative gain; the value 2.4% corresponds instead to comparing 83.43 with the authors' own gpt4-o row (81.48). The relative gain should be computed with respect to an explicitly named baseline, and absolute F1 deltas should be reported alongside relative percentages.
[§4.1 and §4.2] All F1, FP, and FN values come from single runs with no seeds, repeats, confidence intervals, or significance tests. The ID gain is 0.2 F1 points over roughly 1,610 binary examples, and the OOD gains are computed on subsets of 162 to 1,013 examples; at these sample sizes the reported deltas are within the noise of the evaluation. Repeated runs with variance estimates and, where applicable, paired significance tests are needed before the improvements claimed in the abstract can be considered established; otherwise the improvement claims should be withdrawn or substantially tempered.
[§3.2 and Table 3] The attention experiment does not support its stated conclusion. First, no train/test split is described for the fully connected classifiers, and the caption indicates that the evaluation uses the 168 LFQA examples (84 per class); if the classifiers were trained on the same 168 instances, the results are circular. Second, the layers with the highest F1 (layers 1-3, 5-7, 11, 12) report FP near 100 and FN near 0, meaning they assign every instance to the positive class; on balanced classes such a constant-positive predictor attains F1 of about 66.7, which is exactly the reported range of values, so no attention-based signal beyond the class prior is demonstrated. Third, the text's claim of lower false positives than the zero-shot baseline is contradicted by the table: the baseline has FP = 17.85 while most attention rows have FP = 100. Fourth, the abstract and §4.3 state that performance is worse at "layer 4 and layers 8 through 11", but layer 11 has F1 = 66.67, above the baseline F1 = 20; the underperforming layers are 4, 8, 9, and 10. The attention claim should be re-analyzed or removed.

minor comments (6)

[Tables 1 and 2] The tables contain garbled numeric entries (for example, Table 1's gpt4-o row shows "64.71421.2" and Table 2 shows "88.2417.28 7.4"); the tables should be re-typeset and every value checked against the raw model output.
[§4.2] The text states that the method "outperforms the baselines in both ID and OOD sets," but in Table 1 the proposed flan-ul2 row is below the FLAN-UL2 baseline on ExpertQA (55 vs. 59.4); the claim should be restricted to the aggregate averages and to the specific datasets where it holds.
[References] The model citations are imprecise: flan-ul2 should be cited to the UL2 (Tay et al., 2022) and FLAN (Chung et al., 2022) papers rather than to Raffel et al. (2020), and the gpt4-o model is not the GPT-4 model of Achiam et al. (2023); the appropriate model documentation should be cited.
[§3.2] The description does not specify how attention weights are aggregated (over heads, tokens, or positions) before being fed to the fully connected layer; the preprocessing should be stated explicitly.
[§4.1 and Table 3] The FP and FN columns are described in Section 4.1 as rates, but Table 3 reports them as percentages; the units and denominators should be clarified.
[Figure 1] The prompt template asks the model to "Answer the question," but the template contains no question, only a REFERENCE and a CLAIM; the wording should be aligned with the actual prompt.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the zero-shot evaluation is an external benchmark comparison and no fitted parameter is renamed as a prediction.

full rationale

The paper's central claims are empirical benchmark evaluations against the fixed, external AttributionBench labels, not derivations from fitted quantities. The zero-shot textual entailment method uses the prompt in Fig. 1 and reports F1 values; no parameter is fitted to the target result and no equation in the paper reduces to an input. The attention-layer experiment trains or applies per-layer classifiers on attention weights and evaluates F1; although the section omits train/test separation and the reported layer-wise F1 values (e.g., FP=100, FN=0) suggest trivial all-positive classifiers, that is a methodological or correctness weakness, not circularity, because the reported numbers are not equivalent to the method's inputs by construction. The paper contains no self-citations that carry a load-bearing argument, and no uniqueness theorem or ansatz is imported from prior author work. The abstract's percentage improvements are internally inconsistent with Table 1's per-dataset F1s and unreplicated, but arithmetic and replication problems are verification concerns outside the circularity definition. I therefore find no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central numbers are benchmark evaluations, so the ledger is light: no fitted constants enter the zero-shot claim, and no entities are invented. The hand-selected prompt template and the unstated decoding settings are the free choices that determine the reported deltas. The attention half of the paper fits a linear classifier per layer, making those weights fitted parameters, though their values are not reported and the fitted solutions are degenerate constant predictors.

free parameters (2)

Zero-shot prompt template wording
Hand-selected wording in Fig. 1 ('Answer the question with ONLY a YES or NO. Does the REFERENCE entail the CLAIM?') is the sole method change versus the AttributionBench baselines; the headline gains are effects of this template, and decoding settings are not reported.
Per-layer attention classifier weights
Section 3.2 fits a fully connected layer on each layer's attention weights for binary attribution on LFQA; weights are fitted to benchmark data, no train/test split or hyperparameters are reported, and the fitted solutions collapse to constant predictors.

assumptions (4)

domain assumption AttributionBench labels and evaluation protocol are accepted as ground truth for citation reliability.
Section 4.2 adopts AttributionBench's metrics and subsets without re-validating its labels; all headline numbers are F1 on this benchmark.
domain assumption Attribution is validly operationalized as textual entailment between REFERENCE and CLAIM.
Section 3.1 defines the task as Entail(REFERENCE, CLAIM); if entailment is not the right support criterion, the framing measures a different property than attribution.
domain assumption Attention weights from flan-t5-small carry attribution-relevant signal usable by a linear classifier.
Section 3.2 assumes the attention inputs to the linear layer encode the attribution decision; the reported all-positive solutions indicate this assumption produced no informative classifier.
standard math Standard binary classification metrics (F1, FP, FN) are computed as defined in Section 4.1.
Even routine metric computation is in doubt: the stated ID average 73.8 does not match the mean of the listed flan-ul2 F1 values (about 74.9) under the natural table parse.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Document Attribution: Examining Citation Relationships using Large Language Models." pith.science (2026). https://pith.science/paper/LJ73C25Z

@misc{pith2026250506324,
  author       = {Pith},
  title        = {Pith review of: Document Attribution: Examining Citation Relationships using Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LJ73C25Z}},
  note         = {Machine review of arXiv:2505.06324}
}

read the original abstract

As Large Language Models (LLMs) are increasingly applied to document-based tasks - such as document summarization, question answering, and information extraction - where user requirements focus on retrieving information from provided documents rather than relying on the model's parametric knowledge, ensuring the trustworthiness and interpretability of these systems has become a critical concern. A central approach to addressing this challenge is attribution, which involves tracing the generated outputs back to their source documents. However, since LLMs can produce inaccurate or imprecise responses, it is crucial to assess the reliability of these citations. To tackle this, our work proposes two techniques. (1) A zero-shot approach that frames attribution as a straightforward textual entailment task. Our method using flan-ul2 demonstrates an improvement of 0.27% and 2.4% over the best baseline of ID and OOD sets of AttributionBench, respectively. (2) We also explore the role of the attention mechanism in enhancing the attribution process. Using a smaller LLM, flan-t5-small, the F1 scores outperform the baseline across almost all layers except layer 4 and layers 8 through 11.

Figures

Figures reproduced from arXiv: 2505.06324 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 4 canonical work pages

[1]

Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . arXiv preprint arXiv:2303.08774

arXiv 2023
[2]

Tosin Adewumi, Nudrat Habib, Lama Alkhaled, and Elisa Barney. 2024. https://arxiv.org/abs/2404.04631 On the limitations of large language models (llms): False attribution . Preprint, arXiv:2404.04631

arXiv 2024
[3]

Bernd Bohnet, Vinh Q. Tran, Pat Verga, Roee Aharoni, Daniel Andor, Livio Baldini Soares, Massimiliano Ciaramita, Jacob Eisenstein, Kuzman Ganchev, Jonathan Herzig, Kai Hui, Tom Kwiatkowski, Ji Ma, Jianmo Ni, Lierni Sestorain Saralegui, Tal Schuster, William W. Cohen, Michael Collins, Dipanjan Das, Donald Metzler, Slav Petrov, and Kellie Webster. 2023. htt...

arXiv 2023
[4]

Muhammad Khalifa, David Wadden, Emma Strubell, Honglak Lee, Lu Wang, Iz Beltagy, and Hao Peng. 2024. https://openreview.net/forum?id=UPyWLwciYz Source-aware training enables knowledge attribution in language models . In First Conference on Language Modeling

2024
[5]

LLM Attributor: Interactive Visual Attribution for LLM Generation

Seongmin Lee, Zijie J. Wang, Aishwarya Chakravarthy, Alec Helbling, ShengYun Peng, Mansi Phute, Duen Horng Chau, and Minsuk Kahng. 2024. https://arxiv.org/abs/2404.01361 Llm attributor: Interactive visual attribution for llm generation . Preprint, arXiv:2404.01361

work page Pith review arXiv 2024
[6]

Dongfang Li, Zetian Sun, Xinshuo Hu, Zhenyu Liu, Ziyang Chen, Baotian Hu, Aiguo Wu, and Min Zhang. 2023. https://arxiv.org/abs/2311.03731 A survey of large language models attribution . Preprint, arXiv:2311.03731

arXiv 2023
[7]

Yifei Li, Xiang Yue, Zeyi Liao, and Huan Sun. 2024. https://doi.org/10.18653/v1/2024.findings-acl.886 A ttribution B ench: How hard is automatic attribution evaluation? In Findings of the Association for Computational Linguistics: ACL 2024, pages 14919--14935, Bangkok, Thailand. Association for Computational Linguistics

doi:10.18653/v1/2024.findings-acl.886 2024
[8]

Vivek Miglani, Aobo Yang, Aram Markosyan, Diego Garcia-Olano, and Narine Kokhlikyan. 2023. https://doi.org/10.18653/v1/2023.nlposs-1.19 Using captum to explain generative language models . In Proceedings of the 3rd Workshop for Natural Language Processing Open Source Software (NLP-OSS 2023), pages 165--173, Singapore. Association for Computational Linguistics

doi:10.18653/v1/2023.nlposs-1.19 2023

Show all 14 references

[9]

Ramakanth Pasunuru, Koustuv Sinha, Armen Aghajanyan, LILI YU, Tianlu Wang, Daniel M Bikel, Luke Zettlemoyer, Maryam Fazel-Zarandi, and Asli Celikyilmaz. 2023. Eliciting attributions from llms with minimal supervision

2023
[10]

Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. https://jmlr.org/papers/volume21/20-074/20-074.pdf Exploring the limits of transfer learning with a unified text-to-text transformer . Journal of ...

2020
[11]

Xiang Yue, Boshi Wang, Ziru Chen, Kai Zhang, Yu Su, and Huan Sun. 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.307 Automatic evaluation of attribution by large language models . In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 4615--4635...

2023 doi
[12]

Wei Zhou, Heike Adel, Hendrik Schuff, and Ngoc Thang Vu. 2024. https://aclanthology.org/2024.lrec-main.600/ Explaining pre-trained language models with attribution scores: An analysis in low-resource settings . In Proceedings of the 2024 Joint International Conference on Compu...

2024
[13]

online" 'onlinestring :=

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...
[14]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . arXiv preprint arXiv:2303.08774

arXiv 2023

[2] [2]

Tosin Adewumi, Nudrat Habib, Lama Alkhaled, and Elisa Barney. 2024. https://arxiv.org/abs/2404.04631 On the limitations of large language models (llms): False attribution . Preprint, arXiv:2404.04631

arXiv 2024

[3] [3]

Bernd Bohnet, Vinh Q. Tran, Pat Verga, Roee Aharoni, Daniel Andor, Livio Baldini Soares, Massimiliano Ciaramita, Jacob Eisenstein, Kuzman Ganchev, Jonathan Herzig, Kai Hui, Tom Kwiatkowski, Ji Ma, Jianmo Ni, Lierni Sestorain Saralegui, Tal Schuster, William W. Cohen, Michael Collins, Dipanjan Das, Donald Metzler, Slav Petrov, and Kellie Webster. 2023. htt...

arXiv 2023

[4] [4]

Muhammad Khalifa, David Wadden, Emma Strubell, Honglak Lee, Lu Wang, Iz Beltagy, and Hao Peng. 2024. https://openreview.net/forum?id=UPyWLwciYz Source-aware training enables knowledge attribution in language models . In First Conference on Language Modeling

2024

[5] [5]

LLM Attributor: Interactive Visual Attribution for LLM Generation

Seongmin Lee, Zijie J. Wang, Aishwarya Chakravarthy, Alec Helbling, ShengYun Peng, Mansi Phute, Duen Horng Chau, and Minsuk Kahng. 2024. https://arxiv.org/abs/2404.01361 Llm attributor: Interactive visual attribution for llm generation . Preprint, arXiv:2404.01361

work page Pith review arXiv 2024

[6] [6]

Dongfang Li, Zetian Sun, Xinshuo Hu, Zhenyu Liu, Ziyang Chen, Baotian Hu, Aiguo Wu, and Min Zhang. 2023. https://arxiv.org/abs/2311.03731 A survey of large language models attribution . Preprint, arXiv:2311.03731

arXiv 2023

[7] [7]

Yifei Li, Xiang Yue, Zeyi Liao, and Huan Sun. 2024. https://doi.org/10.18653/v1/2024.findings-acl.886 A ttribution B ench: How hard is automatic attribution evaluation? In Findings of the Association for Computational Linguistics: ACL 2024, pages 14919--14935, Bangkok, Thailand. Association for Computational Linguistics

doi:10.18653/v1/2024.findings-acl.886 2024

[8] [8]

Vivek Miglani, Aobo Yang, Aram Markosyan, Diego Garcia-Olano, and Narine Kokhlikyan. 2023. https://doi.org/10.18653/v1/2023.nlposs-1.19 Using captum to explain generative language models . In Proceedings of the 3rd Workshop for Natural Language Processing Open Source Software (NLP-OSS 2023), pages 165--173, Singapore. Association for Computational Linguistics

doi:10.18653/v1/2023.nlposs-1.19 2023