Pith. sign in

REVIEW 5 major objections 5 minor 34 references

NewsEdits 2.0: Learning the Intentions Behind Updating News

T0 review · 5 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that linguistic features in an old news draft predict which facts will later be updated, and that these predictions let an LLM abstain from stale questions almost as well as knowing ground truth.

desk verdict Useful taxonomy and corpus for news edit intentions, but the near-oracle abstention claim rests on a thinly validated GPT-4 judge and small differences. read the letter →

arxiv 2411.18811 v1 pith:PL4AHGB6 submitted 2024-11-27 cs.CL cs.AIcs.DL

Alexander Spangher , Kung-Hsiang Huang , Hyundong Cho , Jonathan May This is my paper · ORCID
classification cs.CLcs.AIcs.DL
keywords newsrevisionhistoryfactualupdatesedit-intentiontaxonomyLLMabstentionquestionansweringdiscoursesilver-labelcorpustemporalreasoning
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

News articles are living documents: as events unfold, facts get corrected, expanded, or replaced, and an LLM answering from an old draft can confidently give outdated answers. This paper tests the hypothesis that the text of an article itself carries linguistic signals of 'factual fluidity' — that a sentence about an advisory or a developing event is more likely to be factually rewritten than a sentence describing a historical detail or a quote. To test this, the authors build the NewsEdits 2.0 taxonomy, a schema of edit intentions drawn up with professional journalists, annotate over 9,200 revision pairs, and train an ensemble model that tags edits at 75.1 micro-F1. Using those tags to silver-label a large revision corpus, they train a predictor of whether a sentence in an older draft will receive a factual update, and find high precision (0.74) in the model's high-confidence region. They then show that when an LLM is given these update-likelihood predictions, it abstains from answering potentially outdated questions at near-oracle accuracy.

What carries the argument

The central object is the NewsEdits 2.0 taxonomy, a four-coarse, twenty-fine-grained schema of edit intentions (Factual, Style, Narrative/Contextual, Other) developed with professional journalists to separate fact updates from stylistic and narrative rewrites. Around that schema the paper builds a pipeline: a sequence-to-sequence LED (Longformer) model tags sentence-revision pairs with edit intentions; its tags silver-label a large corpus; a second Longformer, fed only the old sentence and article, predicts the probability of a factual update; and those probabilities are binned into low/medium/high and injected into an LLM's prompt for a RealTimeQA-style abstention task. The load-bearing mechanism is the two-stage transfer: semantic edit-intention labels learned on a small gold-annotated set are used to train the factual-fluidity predictor at scale, and the predictor's high-precision region drives the downstream abstention gains.

What would settle it

Re-run the abstention experiment with an independently human-annotated set of at least 1,000 question-answer pairs (or with judges from two different LLMs) and compare abstention F1; if the update-prediction variant no longer matches the oracle condition, the near-oracle claim fails. A second check: replace the silver labels used for training the factual-update predictor with gold labels from the annotated test set and see whether the high-precision region and downstream abstention gains persist.

Watch

Extended reading notes

Core claim

The central claim is that factual updates to news sentences are not random: they are signalled by linguistic cues in the original text, and a model trained on revision histories can learn those cues. The paper isolates fact updates from stylistic and narrative edits using the NewsEdits 2.0 edit-intention taxonomy, and shows that pretrained LLMs (GPT-3.5, GPT-4) perform poorly at this prediction task, while a fine-tuned Longformer reaches a macro-F1 of 0.58 overall and precision of 0.74 when predictions are restricted to its most confident sentences. In the paper's use-case, giving GPT-4 a binned update-likelihood score for the retrieved sentence leads to abstention behavior that matches or slightly exceeds the oracle condition, where GPT-4 is told the ground-truth update status.

Load-bearing premise

The load-bearing premise is that the automated judge used to score abstentions—a GPT-4 evaluator validated on only 100 manually labeled questions—correctly decides when a question is answerable from the old sentence and when its answer conflicts with the revised sentence. If that judge is biased, the reported near-oracle abstention accuracy could be inflated.

Editorial extensions

If this is right

  • Retrieval-augmented QA systems could use the update-likelihood signal to abstain or flag uncertainty on sentences likely to go stale, reducing the spread of outdated information.
  • The NewsEdits 2.0 schema and the silver-labeled corpus of 4 million revision pairs provide a reusable resource for studying style and narrative edits, not just factual ones.
  • The finding that sentence-level cues carry most of the predictive signal means lightweight, text-only models can flag fluid facts without external retrieval or knowledge bases.
  • The near-oracle abstention result implies that, in the high-confidence region, a cheap text-only model can stand in for ground-truth update knowledge.

Reading between the lines

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

  • The two-stage pipeline (annotate edit intentions, then predict factual updates) could transfer to other versioned text domains—product documentation, wikis, legal filings—where future changes may also leave linguistic traces; this is an inference, not a paper claim.
  • A newsroom tool could surface only the model's highest-confidence update candidates for human re-verification, a natural productization of the reported high-precision region that the paper does not propose.
  • Because the abstention experiments used a single LLM and a judge validated on only 100 questions, testing the gains across other LLMs and with larger human-judged sets would show whether the near-oracle result generalizes; this is an extension, not a paper claim.
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.

Referee Report

5 major / 5 minor

Summary. This paper introduces NewsEdits 2.0, a taxonomy of edit intentions (Factual, Style, Narrative, Other) with 20 fine-grained categories, developed with professional journalists. The authors annotate 9,200 sentence-revision pairs, train LED-based sequence-to-sequence models to tag edit intentions, then use silver labels from these models to build a large training set for predicting, from an older sentence alone, whether a fact in it will update in a later version. They evaluate this predictor on held-out gold annotations, report a high-precision region at 0.74 precision, and finally test whether these update-likelihood predictions can improve LLM abstention in a RealTimeQA-style setup, reporting near-oracle accuracy.

Significance. The main contributions are substantial: a new edit-intention taxonomy grounded in journalism practice, a 9,200-pair gold annotation with professional journalists, a 4-million-pair silver-labeled corpus, a text-only factual-update prediction task, and a downstream LLM abstention task. If the results hold, they provide evidence for the hypothesis that linguistic features signal factual fluidity, and a practical way to make LLMs more cautious with outdated documents. The pipeline is largely reproducible from public components, and the authors promise code release. However, the empirical support for the two strongest claims ('high precision' and 'near-oracle abstention') is not yet methodologically solid: the precision figure comes from a post hoc subset, and the abstention evaluation rests on a judge whose final decision has not been validated and whose errors could flip the small reported differences.

major comments (5)
  1. [Section 5.1 (Abstention Rate Evaluations)] The final abstain/answer decision is computed as an OR of two GPT-4 judgments: “Is this question answerable given the old sentence?” and “Is the answer consistent with the revised sentence?” The validation on 100 questions reports per-question kappas (>0.74) but not agreement on the derived final decision. Because errors in either sub-judgment propagate through the OR, the small accuracy gaps in Table 8 (e.g., 0.6 points in Maybe-Conflict and 2.4 points in Likely-Conflict between Update Pred and Uniform Warning; Update Pred within 1.6 points of Oracle in every row) are within the likely error band of the judge. The “near-oracle” claim needs human validation of the final decision, ideally with the judge blind to the experimental condition, plus a confusion matrix for the judge-versus-human final labels.
  2. [Tables 4, 8, 9] No confidence intervals or significance tests are reported, and the sizes of the gold test set (Part 2) and the question set (Part 3) are not stated in the main text. The differences that carry the paper’s claims are small — for instance, in Table 8 the Uniform Warning and Update Pred variants differ by only 0.6 points in Maybe-Conflict micro-F1, and Update Pred is essentially tied with Oracle. Without an N, a confidence interval, or a significance test, these point estimates cannot support the claim of near-oracle performance.
  3. [Section 4.2 and Figure 4] The advertised precision of .74 is computed on a post hoc high-confidence subset of sentences. The manuscript does not explain how the confidence threshold is chosen or whether it is tuned on a held-out set. If the threshold is selected after looking at the gold test labels, the .74 is an optimistic estimate, and the abstract’s claim to predict updates “with high precision” should be explicitly scoped to the high-likelihood region rather than to the general prediction task.
  4. [Sections 4.1 and 5.1] The factual-update predictor is trained on silver labels generated by the authors’ own edit-intention model, and the Oracle variant in the Part 3 experiments is defined using the same NewsEdits 2.0 gold taxonomy. Since the Part 2 test labels and the Oracle labels share that taxonomy, systematic biases in the schema (or in the silver-labeling step) are shared across training, evaluation, and the upper-bound condition. This is not a fatal circularity because the Part 2 predictor is evaluated on held-out gold annotations, but the manuscript should discuss this shared-schema limitation and, if possible, include a robustness check using a narrower definition of factual updates (e.g., only Event Update and Correction labels).
  5. [Table 4 and Section 4.2] The FT Longformer Direct Context row reports Macro F1 = 87.8 with class F1s of 22.3 (Fact) and 93.0 (Not Fact), which is arithmetically inconsistent (the average of the two class F1s is about 57.7). This looks like a typo, but it prevents the reader from trusting the table. In addition, the text “At 41.2 F1-score, the journalist sets a moderately higher upper bound” should identify that 41.2 is the Fact-class F1, not the macro F1; used without this clarification it understates the model’s macro F1 of 58.0.
minor comments (5)
  1. [Abstract and Section 4.1] There are typos: “absention” in the abstract and “acnkowledge” in Section 4.1; please correct them.
  2. [Table 8] The “Avg.” column is identical to the Micro F1 column in every row; please state what it averages or remove it.
  3. [Section 3.2] The final inter-annotator agreement on the full 9,200-pair annotation is not reported; only the training criterion (kappa > .6 on a 50-pair gold set) is given. Please report the kappa or other agreement measure on the final annotated data.
  4. [Section 5.1] The number of sentence pairs and the number of questions used in the LLM-QA experiment are not reported in the main text; please add these counts, as they are needed to assess the reliability of the point estimates in Tables 8 and 9.
  5. [Appendix D.3] The evaluation prompt instructs the judge to answer question 2 with n/a when question 1 is “no”, but the main text says to abstain if either answer is “no”; please specify how n/a is mapped to the final abstain/answer label.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the factual-update predictor is trained on silver labels but tested on independent gold annotations, and the abstention claim is a direct comparison against an oracle signal.

full rationale

The paper's derivation chain is not circular in any step that can be pinned to a definitional equivalence or a fitted value renamed as a prediction. Part 2 trains a factual-update predictor on silver labels produced by the Part 1 edit-intention model: the paper states, 'Then, we use the best-performing edit-intentions model, in Section 3.3, to produce silver labels. We assign labels l using both versions of a revision pair (Equation 1); then we discard D′, s′j and try to predict l using just D, si (Equation 2).' This is a standard silver-labeling setup, not a reduction: the predictor must learn a function from the old sentence alone, and it is evaluated on the independently gold-annotated test set ('We test all approaches on the same gold-labeled documents Dgold test , which were part of our gold-annotated test set'). The reported macro-F1 of .58 and the high-precision subset precision of .74 are measured against that gold set, so the central prediction claim is not forced by construction. Part 3's abstention claim is an empirical comparison between the paper's predicted update likelihoods and oracle gold labels in the same LLM prompting framework; the paper reports that the predicted variant scores 90.6 vs. 90.0 micro-F1 on Maybe-Conflict and 67.1 vs. 66.5 on Likely-Conflict, which are close but not identical to the oracle, showing the comparison is not a tautology. The GPT-4-based abstention judge is a validity concern, and its 100-question validation is thin, but the judge does not consume the update predictions when scoring, so any bias is a measurement issue rather than circularity. The many self-citations to NewsEdits and to the authors' discourse, quote, and source-detection models are citations to datasets and feature extractors whose outputs are empirically combined and ablated; they are not invoked as unverified uniqueness theorems or ansatze that force the paper's conclusions. No equation in the paper reduces to another equation by definition, and no fitted parameter is renamed as a prediction. The central empirical claims therefore stand on independent gold evaluation and a direct oracle comparison.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claims rest on the quality of the authors' annotated taxonomy, the accuracy of model-generated silver labels used as training signal, and the validity of a GPT-4-based evaluation judge. The hand-chosen sampling thresholds and high-confidence cutoff are free parameters that shape the reported precision.

free parameters (5)
  • Dataset update thresholds = >10% added, >5% deleted
    Section 4.1 selects 500,000 articles with these thresholds; hand-chosen to focus on articles likely to update, introducing selection bias.
  • High-precision region threshold = not specified
    Section 4.2 and Figure 4 report 0.74 precision on a subset of sentences the model is most confident about; the cutoff is not defined in the main text.
  • Low/medium/high update-likelihood bins = not specified
    Section 5.1 bins model scores into three categories for the LLM prompt; bin boundaries are not given.
  • Training class balance = 50/50
    Section 4.2 balances the training dataset to equal class counts, changing the prior and affecting absolute F1 values.
  • Annotator agreement threshold = kappa > 0.6
    Section 3.2 uses this hand-set agreement level as a quality gate for annotators.
assumptions (5)
  • domain assumption NewsEdits sentence alignments across revisions are mostly correct.
    Sections 3.1 and A.2 define edit intentions on linked sentence pairs; the paper improves linking to 95 F1, but residual matching errors propagate into labels.
  • domain assumption Silver labels from the LED model are accurate enough for training the factual-update predictor.
    Section 4.1 uses model-generated labels as all training supervision for Part 2; if the tagger is systematically biased, the predictor may not generalize.
  • domain assumption The NewsEdits 2.0 taxonomy captures the real reasons journalists edit.
    Section 3.1 develops the schema with journalists, but there is no independent validation of the schema itself against an external standard.
  • domain assumption GPT-4 judge agreement on 100 examples generalizes to the full test set.
    Section 5.1 validates the evaluation procedure on only 100 manually labeled questions before applying it to the full abstention evaluation.
  • domain assumption Linguistic cues such as temporality, statistics, and developing-event language are stable across news domains.
    Section 4.2 and Table 5 rely on manual annotation of 100 sentences; the cue distribution may shift across topics or outlets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NewsEdits 2.0: Learning the Intentions Behind Updating News." pith.science (2026). https://pith.science/paper/PL4AHGB6

@misc{pith2026241118811,
  author       = {Pith},
  title        = {Pith review of: NewsEdits 2.0: Learning the Intentions Behind Updating News},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PL4AHGB6}},
  note         = {Machine review of arXiv:2411.18811}
}
read the original abstract

As events progress, news articles often update with new information: if we are not cautious, we risk propagating outdated facts. In this work, we hypothesize that linguistic features indicate factual fluidity, and that we can predict which facts in a news article will update using solely the text of a news article (i.e. not external resources like search engines). We test this hypothesis, first, by isolating fact-updates in large news revisions corpora. News articles may update for many reasons (e.g. factual, stylistic, narrative). We introduce the NewsEdits 2.0 taxonomy, an edit-intentions schema that separates fact updates from stylistic and narrative updates in news writing. We annotate over 9,200 pairs of sentence revisions and train high-scoring ensemble models to apply this schema. Then, taking a large dataset of silver-labeled pairs, we show that we can predict when facts will update in older article drafts with high precision. Finally, to demonstrate the usefulness of these findings, we construct a language model question asking (LLM-QA) abstention task. We wish the LLM to abstain from answering questions when information is likely to become outdated. Using our predictions, we show, LLM absention reaches near oracle levels of accuracy.

Figures

Figures reproduced from arXiv: 2411.18811 by the authors.

Figure 1
Figure 1. Updates can occur for many different reasons. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall paper flow. In Part 1 of our paper, we develop an edits-intention scheme to describe news edits and train models to apply this schema to existing news revision corpora (Spangher et al., 2022). In Part 2, we use these models to silver-label a large corpus and ask how well we can predict whether a sentence will factually update. In Part 3, we show these predictions can be beneficial for increasing abstention r… view at source ↗
Figure 3
Figure 3. NewsEdits 2.0: Edit-Intentions Schema categories and their subcategories. In this work, we focus mainly on the Factual Edit category. See Appendix C.1 for definitions for all categories. ing similar schemas to what we have developed. While building NewsEdits 2.0, we were inspired by the schemas developed by prior work and they pro￾vided a starting point for our taxonomy. We added edit-categories that were more journ… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Performance of Fact-update model increases [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: The portion of annotation tasks assigned to [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Coarse-grained confusion matrix for the LED [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Fine-grained confusion matrix for the LED model trained with Discourse and Argumentation features. [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]
Figure 8
Figure 8. Figure 8: The interface for annotating edit intentions. [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 20 canonical work pages

  1. [1]

    Khalid Al-Khatib, Henning Wachsmuth, Johannes Kiesel, Matthias Hagen, and Benno Stein. 2016. https://aclanthology.org/C16-1324/ A News Editorial Corpus for Mining Argumentation Strategies . In 26th International Conference on Computational Linguistics (COLING 2016), pages 3433--3443. Association for Computational Linguistics

    work page 2016

  2. [2]

    Iz Beltagy, Matthew E Peters, and Arman Cohan. 2020. Longformer: The long-document transformer. arXiv preprint arXiv:2004.05150

    arXiv 2020

  3. [3]

    Prafulla Kumar Choubey, Aaron Lee, Ruihong Huang, and Lu Wang. 2020. Discourse as a function of event: Profiling discourse structure in news articles around the main event. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics

    work page 2020

  4. [4]

    Sarah Cohen, James T Hamilton, and Fred Turner. 2011. Computational journalism. Communications of the ACM, 54(10):66--71

    work page 2011

  5. [5]

    H.D. Croly. 1943. https://books.google.com/books?id=cDgQAAAAIAAJ The New Republic . v. 108. Republic Publishing Company

    work page 1943

  6. [6]

    Ido Dagan, Oren Glickman, and Bernardo Magnini. 2005. The pascal recognising textual entailment challenge. In Machine learning challenges workshop, pages 177--190. Springer

    2005

  7. [7]

    George R Doddington, Alexis Mitchell, Mark A Przybocki, Lance A Ramshaw, Stephanie M Strassel, and Ralph M Weischedel. 2004. The automatic content extraction (ace) program-tasks, data, and evaluation. In Lrec, volume 2, pages 837--840. Lisbon

    work page 2004

  8. [8]

    Manaal Faruqui, Ellie Pavlick, Ian Tenney, and Dipanjan Das. 2018. Wikiatomicedits: A multilingual corpus of wikipedia edits for modeling language and discourse. arXiv preprint arXiv:1808.09422

    arXiv 2018

Show all 34 references
  1. [9]

    Haleluya Hadero and David Bauder. 2023. New york times sues microsoft, open ai over use of content. Globe & Mail (Toronto, Canada), pages B1--B1

    2023

  2. [10]

    Rujun Han, Xiang Ren, and Nanyun Peng. 2020. Econet: Effective continual pretraining of language models for event temporal reasoning. arXiv preprint arXiv:2012.15283

    2020 arXiv

  3. [11]

    I Hsu, Kuan-Hao Huang, Elizabeth Boschee, Scott Miller, Prem Natarajan, Kai-Wei Chang, Nanyun Peng, et al. 2021. Degree: A data-efficient generation-based event extraction model. arXiv preprint arXiv:2108.12724

    2021 arXiv

  4. [12]

    Kung-Hsiang Huang, Sam Tang, and Nanyun Peng. 2021. https://doi.org/10.18653/v1/2021.emnlp-main.426 Document-level entity-based extraction as template generation . In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 5257--5269, Onli...

    2021 doi

  5. [13]

    Zhen Jia, Abdalghani Abujabal, Rishiraj Saha Roy, Jannik Str \"o tgen, and Gerhard Weikum. 2018. Tempquestions: A benchmark for temporal question answering. In Companion Proceedings of the The Web Conference 2018, pages 1057--1062

    2018

  6. [14]

    Xiaoqi Jiao, Yichun Yin, Lifeng Shang, Xin Jiang, Xiao Chen, Linlin Li, Fang Wang, and Qun Liu. 2019. Tinybert: Distilling bert for natural language understanding. arXiv preprint arXiv:1909.10351

    2019 arXiv

  7. [15]

    Jungo Kasai, Keisuke Sakaguchi, Yoichi Takahashi, Ronan Le Bras, Akari Asai, Xinyan Yu, Dragomir Radev, Noah A Smith, Yejin Choi, and Kentaro Inui. 2022. Realtime qa: What's the answer right now? arXiv preprint arXiv:2207.13332

    2022 arXiv

  8. [16]

    Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, and Radu Soricut. 2019. Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:1909.11942

    2019 arXiv

  9. [17]

    Sha Li, Heng Ji, and Jiawei Han. 2021. https://doi.org/10.18653/v1/2021.naacl-main.69 Document-level event argument extraction by conditional generation . In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Huma...

    2021 doi

  10. [18]

    Adam Liska, Tom \'a s Kocisk \`y , Elena Gribovskaya, Tayfun Terzi, Eren Sezener, and Devang Agrawal. 2022. Cyprien de masson d’autume, tim scholtes, manzil zaheer, susannah young, ellen gilsenan-mcmahon, sophia austin, phil blunsom, and angeliki lazaridou. 2022. streamingqa: ...

    2022

  11. [19]

    Yixin Nie, Adina Williams, Emily Dinan, Mohit Bansal, Jason Weston, and Douwe Kiela. 2020. Adversarial NLI : A new benchmark for natural language understanding. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Association for Computat...

    2020

  12. [20]

    Silvia Pareti, Tim O’keefe, Ioannis Konstas, James R Curran, and Irena Koprinska. 2013. Automatically detecting and attributing indirect quotations. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 989--999

    2013

  13. [21]

    Kostas Saltzis. 2012. Breaking news online: How news stories are updated and maintained around-the-clock. Journalism practice, 6(5-6):702--710

    2012

  14. [22]

    Alexander Spangher, Jonathan May, Sz-Rung Shiang, and Lingjia Deng. 2021. Multitask semi-supervised learning for class-imbalanced discourse classification. In Proceedings of the 2021 conference on empirical methods in natural language processing, pages 498--517

    2021

  15. [23]

    Alexander Spangher, Nanyun Peng, Jonathan May, and Emilio Ferrara. 2023. Identifying informational sources in news articles. arXiv preprint arXiv:2305.14904

    2023 arXiv

  16. [24]

    Alexander Spangher, Xiang Ren, Jonathan May, and Nanyun Peng. 2022. Newsedits: A news article revision dataset and a novel document-level reasoning challenge. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: ...

    2022

  17. [25]

    Alexander Spangher, James Youn, Matthew Debutts, Nanyun Peng, and Jonathan May. 2024. Explaining mixtures of sources in news articles

    2024

  18. [26]

    Qingyu Tan, Hwee Tou Ng, and Lidong Bing. 2023. Towards benchmarking and improving the temporal reasoning capability of large language models. arXiv preprint arXiv:2306.08952

    2023 arXiv

  19. [27]

    Teun A Van Dijk. 1998. News as discourse. Lawrence Erlbaum Associates

    1998

  20. [28]

    Kristian Woodsend and Mirella Lapata. 2011. Wikisimple: Automatic simplification of wikipedia articles. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 25, pages 927--932

    2011

  21. [29]

    Siheng Xiong, Ali Payani, Ramana Kompella, and Faramarz Fekri. 2024. Large language models can learn temporal reasoning. arXiv preprint arXiv:2401.06853

    2024 arXiv

  22. [30]

    Diyi Yang, Aaron Halfaker, Robert Kraut, and Eduard Hovy. 2017. Identifying semantic edit intentions from revisions in wikipedia. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2000--2010

    2017

  23. [31]

    W Victor Yarlott, Cristina Cornelio, Tian Gao, and Mark Finlayson. 2018. Identifying the discourse function of news article paragraphs. In Proceedings of the Workshop Events and Stories in the News 2018, pages 25--33

    2018

  24. [32]

    Fan Zhang and Diane Litman. 2015. Annotation and classification of argumentative writing revisions. In Proceedings of the tenth workshop on innovative use of NLP for building educational applications, pages 133--143

    2015

  25. [33]

    URL: " 'urlintro :=

    ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before...

  26. [34]

    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 12, 2026 · model on record in the stance chip above.