Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Enhancing LLMs for Governance with Human Oversight: Evaluating and Aligning LLMs on Expert Classification of Climate Misinformation for Detecting False or Misleading Claims about Climate Change

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Fine-tuned GPT-3.5-turbo classifies climate misinformation at expert-level agreement.

desk verdict A promising fine-tuning result buried under an unreliable open-source comparison; the expert-alignment result is suggestive but needs harder evidence. read the letter →

arxiv 2501.13802 v2 pith:TTI33TEZ submitted 2025-01-23 cs.CY

Mowafak Allaham , Ayse D. Lokmanoglu , P. Sol Hart , Erik C. Nisbet This is my paper · ORCID
classification cs.CY
keywords climatemisinformationlargelanguagemodelsfine-tuningexpertannotationhumanoversightcontentmoderationAIgovernanceCARDStaxonomy
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

The paper asks whether LLMs can help govern climate misinformation rather than amplify it. The authors benchmark 16 open- and closed-source models on the expert-annotated CARDS test set, then fine-tune GPT-3.5-turbo on the same expert data and test it on 914 paragraphs drawn from low-credibility climate articles shared on Facebook and X in 2022. The central result is that the fine-tuned model labels false or misleading climate claims with the same measured reliability as two senior climate-communication experts: Krippendorff's $\alpha = 0.89$ for both the expert pair and the model against their labels, with a macro F1 of 0.88. On the same sample it clearly outperforms the specialist CARDS RoBERTa model ($\alpha = 0.66$, macro F1 0.74), and on the benchmark it also beats larger proprietary zero-shot models. The authors take this as evidence that expert-annotated training data, as a form of human oversight, is what makes LLMs usable for governance tasks such as content moderation at scale.

What carries the argument

The load-bearing mechanism is the CARDS taxonomy, a set of 27 false-or-misleading claim labels grouped under five super-claims about climate change, together with the expert-annotated paragraphs used to train and evaluate it. The authors convert the CARDS coding manual into prompts for zero-shot evaluation, then fine-tune GPT-3.5-turbo on the CARDS training set using system, user, and assistant message templates so the model emits one of the 27 labels. On the social-media sample, alignment with expert annotations is measured with Krippendorff's $\alpha$ and macro-averaged precision, recall, and F1 at the super-claim level, with the reconciled labels of two senior experts serving as ground truth.

What would settle it

Have an independent panel of climate-communication experts, none involved in this study, code the same 914 paragraphs from the raw CARDS rubric and compute Krippendorff's $\alpha$ between the fine-tuned model and the panel. If that agreement falls well below 0.89, or below the panel's own intercoder agreement, the expert-equivalence claim fails; a complementary check would add health-impact claims outside the CARDS taxonomy, where the paper already reports the model misses many cases.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a relatively small proprietary model fine-tuned on a public expert-annotated taxonomy can classify false or misleading climate claims in wild social-media text at the same level of agreement as senior experts. The fine-tuned GPT-3.5-turbo reaches Krippendorff's $\alpha = 0.89$ against the reconciled expert labels on 914 paragraphs, matching the two experts' intercoder agreement of $\alpha = 0.89$, and a macro F1 of 0.88 at the super-claim level. This exceeds the CARDS RoBERTa model on the same paragraphs ($\alpha = 0.66$, macro F1 0.74) and exceeds GPT-4o and RoBERTa on the original CARDS test set (macro F1 0.84 vs. 0.75 and 0.77). The authors also find that open-source 2B–9B models lag far behind in zero-shot classification, and that existing expert-trained tools outperform most proprietary zero-shot models, including GPT-4o.

Load-bearing premise

The load-bearing assumption is that the reconciled labels of the two senior experts are valid ground truth for what counts as false or misleading climate claims, rather than a local coding preference; if those labels are systematically tied to the same taxonomy and training examples the model was fine-tuned on, the reported expert-level agreement could overstate the model's ability to match independent expert judgment.

Editorial extensions

If this is right

  • Fine-tuning a mid-size proprietary LLM on an expert taxonomy can produce a claim classifier whose agreement with experts on in-the-wild social media text matches, rather than merely approaches, the experts' own intercoder agreement.
  • Expert-trained specialist tools remain strong baselines: CARDS outperforms most proprietary zero-shot models, including GPT-4o, on the benchmark, so domain-specific training data still matters.
  • Open-source models in the 2B–9B parameter range are far behind on this classification task, implying that resource-constrained civil-society organizations would currently need API access or better open training data to use LLMs for this governance task.
  • The paper's own limitation sections imply that expert-level performance on low-credibility sources should not be assumed for high-credibility sources, claims outside the CARDS taxonomy, non-English text, or health-impact claims, where the model misses many cases.

Reading between the lines

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

  • Extension: because the ground-truth labels and the fine-tuning data come from the same CARDS rubric and two of the expert coders are also authors, the reported $\alpha = 0.89$ is best read as alignment with this specific expert coding scheme; an independent coding panel could yield a lower number.
  • Extension: if the fine-tuning recipe transfers, the same human-oversight pattern could be applied to other governance domains such as health or political misinformation, provided an expert-annotated taxonomy exists for each domain.
  • Extension: a useful follow-up would be per-claim calibration on rare sub-claims such as species and plant impacts, where the paper reports systematic misfires; class-imbalance correction or targeted additional expert data would be a direct test of whether the bias is in the training distribution rather than the model.
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

4 major / 5 minor

Summary. This paper evaluates LLMs for classifying false or misleading climate claims. Using the CARDS test set (N=2,904), it compares zero-shot proprietary and open-source LLMs against the RoBERTa-based CARDS model, then fine-tunes GPT-3.5-turbo on CARDS training data and evaluates it both on that test set and on a curated social-media sample (914 paragraphs) labeled by two climate-communication experts. The paper reports that (i) open-source models substantially underperform proprietary models, (ii) CARDS outperforms several proprietary models, and (iii) fine-tuned GPT-3.5-turbo matches expert-level agreement on the social-media sample (macro F1=0.88, Krippendorff's alpha=0.89). It argues that this demonstrates the value of expert-annotated data and human oversight for LLM-based governance tools.

Significance. If fully supported, the paper would provide a useful demonstration that fine-tuning a small, inexpensive proprietary model on expert-annotated data can outperform larger general-purpose models and a task-specific BERT model on an external test set, and can approach expert-level coding on out-of-distribution social-media text. The use of the external CARDS held-out test set for the fine-tuning evaluation, the explicit comparison to expert inter-coder reliability, and the release of the expert-coded sample are strengths that make the core fine-tuning result plausible and non-circular. However, the two headline comparisons—open-source versus proprietary underperformance and expert-level equivalence—currently rest on methodological choices that need repair. With those repairs, the paper could be a valuable case study for human-oversight alignment in governance applications.

major comments (4)
  1. [§5.1, Table 1, Figure 2] Replacing every invalid or unparseable LLM output with a label drawn uniformly at random from 28 possible labels means that models with high non-compliance rates (Figure 2) are scored near a random baseline. The F1 values for open-source models in Table 1 therefore conflate instruction-following with classification skill, and the paper's first contribution—that open-source models substantially underperform—is not established as stated. Please report valid-response rates per model and either (a) evaluate on valid responses only, (b) introduce an explicit 'unparseable/refusal' class, or (c) model abstention separately; in all cases, report uncertainty around the metrics.
  2. [Appendix A.8] The prompt template for open-source models contains a corrupted taxonomy: category '2_1' is defined as 'They changed the name from global warming to climate change', which duplicates '1_8', while the actual CARDS category 'It's natural cycles/variation' is absent. Open-source valid responses were therefore generated under a different label space than the proprietary models and the CARDS ground truth, independently compromising the RQ1/RQ2 comparison even for compliant outputs. The open-source evaluation must be rerun with a corrected prompt, or the paper should report results only for a corrected subset.
  3. [§4.4, §5.2, Table 2] The expert labels used as ground truth for the social-media evaluation were produced by two of the paper's authors applying the same CARDS taxonomy used to fine-tune the model, and the final labels are a reconciled consensus rather than independent codings. The reported model-expert agreement (alpha=0.89, F1=0.88) is therefore not directly comparable to the reported inter-expert alpha=0.89, which measures agreement between two independent codings; a model can match a consensus label more easily than it can match either expert. Please report agreement with each expert separately, state whether annotators were blind to model outputs, and discuss the shared-taxonomy and author-labeling threat to the RQ4 claim of expert-level performance.
  4. [§4.1, Appendices A.4 and A.8] The open-source and proprietary models were evaluated with different prompt templates (the full coding-manual JSON prompt versus the shortened classes prompt), different inference temperatures (0 versus 0.001), and different decoding paths. The reported substantial gap in Table 1 cannot be cleanly attributed to model family or openness because prompt format and compliance are confounded with model choice. The paper should use an as-identical-as-possible evaluation protocol, or explicitly test whether the open-source results change under the proprietary-style prompt.
minor comments (5)
  1. [§5.1 and §4.3] The paper refers to '28 possible labels' (Section 5.1) and '27 false or misleading claim labels' (Section 4.3), but the taxonomy in Appendices A.4 and A.8 contains 26 claim labels plus the no-claim class; please correct these counts.
  2. [Table 1] Please clarify whether the macro-averaged F1 includes the no-claim class; Table 1 says 'five categories of super-claims' while Table 2 includes a no-claim category, and the two tables should use a consistent evaluation scope.
  3. [Throughout] There are numerous typos, including 'reproducability' (§4.1), 'intercoder-relability' (§5.2), 'Krippendorf' (§5.2), and 'Videgen' (reference [46]); please proofread carefully.
  4. [Appendix A.11] The alignment and misalignment symbols in the table appear to be missing or not rendered, making the illustrative examples hard to interpret; please ensure the glyphs display correctly.
  5. [Table 1] The performance differences among the proprietary models (e.g., GPT-4o 0.75 versus GPT-4 0.74) are likely within noise; reporting bootstrap confidence intervals would make the comparison more informative.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central evaluations use an external held-out benchmark and newly collected expert labels; remaining issues are validity and independence concerns, not circular reasoning.

full rationale

The paper's derivation chain is not circular. The CARDS test set (N=2,904) is an external, held-out benchmark that was not used for fine-tuning, so the Table 1/Table 4 F1 comparisons and the fine-tuned GPT-3.5-turbo versus CARDS RoBERTa comparison on that test set are genuine out-of-sample measurements. The social-media evaluation in Section 5.2 compares the fine-tuned model against 914 newly annotated paragraphs that were not in the training data; the expert labels were produced after model inference using the same CARDS taxonomy, but agreement with those labels is not statistically forced by any fitted parameter. Two concerns are noted but do not amount to circularity. First, Section 5.1's replacement of invalid LLM responses with random labels conflates non-compliance with classification error and makes the open-source versus proprietary gap partly a protocol artifact; this is a validity threat to RQ1/RQ2, not a circular derivation. Second, the expert coders may overlap with the authors and applied the same taxonomy used for fine-tuning, which limits the independence of the 'expert equivalence' claim; however, the model never saw the 914 expert labels during training, so the agreement is not definitionally forced. The paper also explicitly acknowledges the single-taxonomy limitation in its Limitations section. No load-bearing self-citation or imported uniqueness theorem is present.

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

The central claims rest on the validity of the CARDS taxonomy, the unbiasedness of the author-provided expert labels, the fairness of the invalid-response replacement procedure, and the representativeness of the 914-paragraph sample. None of these is independently verified, and the first two are partially acknowledged as limitations.

assumptions (4)
  • domain assumption The CARDS taxonomy of 5 super-claims and 27 sub-claims is a complete and valid codification of false/misleading climate claims for both contrarian blogs and low-credibility social media.
    State as ground truth and used for training, prompting, and expert annotation; the paper's Limitations notes the taxonomy misses health-related claims, so this assumption is partially acknowledged.
  • domain assumption The two experts' reconciled labels on 914 paragraphs are an unbiased gold standard.
    The annotators are likely the paper's authors (Hart and Nisbet), which is not flagged; their labels may share coding conventions with the CARDS training data used to fine-tune the model.
  • ad hoc to paper Replacing invalid LLM outputs with uniformly random labels yields a valid performance estimate.
    Section 5.1: 'we replaced invalid responses from LLMs with values randomly sampled from a total of 28 possible labels'; this is not a standard evaluation practice and penalizes prompt non-compliance as if it were error.
  • domain assumption The stratified sample of 914 paragraphs represents the 856,722 paragraph population.
    Half drawn from 'no claim' and half stratified across claim types; small size limits precision.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing LLMs for Governance with Human Oversight: Evaluating and Aligning LLMs on Expert Classification of Climate Misinformation for Detecting False or Misleading Claims about Climate Change." pith.science (2026). https://pith.science/paper/TTI33TEZ

@misc{pith2026250113802,
  author       = {Pith},
  title        = {Pith review of: Enhancing LLMs for Governance with Human Oversight: Evaluating and Aligning LLMs on Expert Classification of Climate Misinformation for Detecting False or Misleading Claims about Climate Change},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TTI33TEZ}},
  note         = {Machine review of arXiv:2501.13802}
}
read the original abstract

Climate misinformation is a problem that has the potential to be substantially aggravated by the development of Large Language Models (LLMs). In this study we evaluate the potential for LLMs to be part of the solution for mitigating online dis/misinformation rather than the problem. Employing a public expert annotated dataset and a curated sample of social media content we evaluate the performance of proprietary vs. open source LLMs on climate misinformation classification task, comparing them to existing climate-focused computer-assisted tools and expert assessments. Results show (1) open-source models substantially under-perform in classifying climate misinformation compared to proprietary models, (2) existing climate-focused computer-assisted tools leveraging expert-annotated datasets continues to outperform many of proprietary models, including GPT-4o, and (3) demonstrate the efficacy and generalizability of fine-tuning GPT-3.5-turbo on expert annotated dataset in classifying claims about climate change at the equivalency of climate change experts with over 20 years of experience in climate communication. These findings highlight 1) the importance of incorporating human-oversight, such as incorporating expert-annotated datasets in training LLMs, for governance tasks that require subject-matter expertise like classifying climate misinformation, and 2) the potential for LLMs in facilitating civil society organizations to engage in various governance tasks such as classifying false or misleading claims in domains beyond climate change such as politics and health science.

Figures

Figures reproduced from arXiv: 2501.13802 by the authors.

Figure 1
Figure 1. Taxonomy of false or misleading claims published by Coan et al. [15] [PITH_FULL_IMAGE:figures/full_fig_p015_1.png] view at source ↗
Figure 2
Figure 2. Proportion of valid responses for each LLM in the zero-shot task described in section 4.1 to classify false or misleading claims about climate change in the CARDS test dataset. Numbers inside the bar plot represent the number of invalid responses by each model [PITH_FULL_IMAGE:figures/full_fig_p022_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Informing AI Risk Assessment with News Media: Analyzing National and Political Variation in the Coverage of AI Risks

    cs.CY 2025-07 conditional novelty 6.0 of 10

    AI risk coverage in the news varies by country and by U.S. outlet political bias, with right-leaning outlets emphasizing malicious actors and political-culture risks.

Reference graph

Works this paper leans on

48 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [1]

    D. A. Scheufele, Science communication as political communication, Proceedings of the National Academy of Sciences 111 (2014) 13585–13592

    work page 2014

  2. [2]

    J. Allgaier, Science and environmental communication on youtube: Strategically distorted commu- nications in online videos on climate change and climate engineering, Frontiers in communication 4 (2019) 446007

    work page 2019

  3. [3]

    Gounaridis, J

    D. Gounaridis, J. P. Newell, The social anatomy of climate change denial in the united states, Scientific Reports 14 (2024) 2097

    work page 2024

  4. [4]

    IPCC, Climate Change 2022: Impacts, Adaptation and Vulnerability, Summary for Policymakers, Cambridge University Press, Cambridge, UK and New York, USA, 2022

    work page 2022

  5. [5]

    Lewandowsky, Climate change disinformation and how to combat it, Annual Review of Public Health 42 (2021) 1–21

    S. Lewandowsky, Climate change disinformation and how to combat it, Annual Review of Public Health 42 (2021) 1–21

    work page 2021

  6. [6]

    K. M. d. Treen, H. T. Williams, S. J. O’Neill, Online misinformation about climate change, Wiley Interdisciplinary Reviews: Climate Change 11 (2020) e665

    work page 2020

  7. [7]

    M. Fore, S. Singh, C. Lee, A. Pandey, A. Anastasopoulos, D. Stamoulis, Unlearning climate misinformation in large language models, arXiv preprint arXiv:2405.19563 (2024)

    arXiv 2024

  8. [8]

    Zhang, K

    Y. Zhang, K. Sharma, L. Du, Y. Liu, Toward mitigating misinformation and social media manip- ulation in llm era, in: Companion Proceedings of the ACM on Web Conference 2024, 2024, pp. 1302–1305

    work page 2024

Show all 48 references
  1. [9]

    Climate Action Against Disinformation’s report highlights how platforms have responded to the EU legislation for online safety so far

    CAAD, Underperforming & unprepared, 2024. Climate Action Against Disinformation’s report highlights how platforms have responded to the EU legislation for online safety so far

    2024

  2. [10]

    Ellison, B

    T. Ellison, B. Hugh, Climate security and misinformation: A baseline, 2024

    2024

  3. [11]

    Kreps, R

    S. Kreps, R. M. McCain, M. Brundage, All the news that’s fit to fabricate: Ai-generated text as a tool of media misinformation, Journal of experimental political science 9 (2022) 104–117

    2022

  4. [12]

    Marlow, S

    T. Marlow, S. Miller, J. T. Roberts, Twitter discourses on climate change: exploring topics and the presence of bots (2020)

    2020

  5. [13]

    CCDH, The toxic ten: How 10 fringe publishers fuel 69% of digital climate change denial, 2021. Report

    2021

  6. [14]

    Romero-Vicente, Platforms’ policies on climate change misinformation, 2023

    A. Romero-Vicente, Platforms’ policies on climate change misinformation, 2023. Factsheet

    2023

  7. [15]

    T. G. Coan, C. Boussalis, J. Cook, M. O. Nanko, Computer-assisted classification of contrarian claims about climate change, Scientific reports 11 (2021) 22320

    2021

  8. [16]

    H. T. Vu, A. Baines, N. Nguyen, Fact-checking climate change: An analysis of claims and verification practices by fact-checkers in four countries, Journalism & Mass Communication Quarterly 100 (2023) 286–307

    2023

  9. [17]

    Leippold, S

    M. Leippold, S. A. Vaghefi, D. Stammbach, V. Muccione, J. Bingler, J. Ni, C. Colesanti-Senni, T. Wekhof, T. Schimanski, G. Gostlow, et al., Automated fact-checking of climate change claims with large language models, arXiv preprint arXiv:2401.12566 (2024)

    2024 arXiv

  10. [18]

    Stiff, F

    H. Stiff, F. Johansson, Detecting computer-generated disinformation, International Journal of Data Science and Analytics 13 (2022) 363–383

    2022

  11. [19]

    J. Ni, F. Xue, X. Yue, Y. Deng, M. Shah, K. Jain, G. Neubig, Y. You, Mixeval: Deriving wisdom of the crowd from llm benchmark mixtures, arXiv preprint arXiv:2406.06565 (2024)

    2024 arXiv

  12. [20]

    Thulke, Y

    D. Thulke, Y. Gao, P. Pelser, R. Brune, R. Jalota, F. Fok, M. Ramos, I. van Wyk, A. Nasir, H. Goldstein, T. Tragemann, K. Nguyen, A. Fowler, A. Stanco, J. Gabriel, J. Taylor, D. Moro, E. Tsymbalov, J. de Waal, E. Matusov, M. Yaghi, M. Shihadah, H. Ney, C. Dugast, J. Dotan, D. ...

    2024 arXiv

  13. [21]

    Lacombe, K

    R. Lacombe, K. Wu, E. Dilworth, Climatex: Do llms accurately assess human expert confidence in climate statements?, arXiv preprint arXiv:2311.17107 (2023)

    2023 arXiv

  14. [22]

    Gehrmann, E

    S. Gehrmann, E. Clark, T. Sellam, Repairing the cracked foundation: A survey of obstacles in evaluation practices for generated text, Journal of Artificial Intelligence Research 77 (2023) 103–166

    2023

  15. [23]

    Z. Xiao, W. H. Deng, M. S. Lam, M. Eslami, J. Kim, M. Lee, Q. V. Liao, Human-centered evaluation and auditing of language models, in: Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, 2024, pp. 1–6

    2024

  16. [24]

    K. Yang, F. Menczer, Anatomy of an ai-powered malicious social botnet, Journal of Quantitative Description: Digital Media 4 (2024). URL: http://dx.doi.org/10.51685/jqd.2024.icwsm.7. doi: 10. 51685/jqd.2024.icwsm.7

    2024 doi

  17. [25]

    Ferrara, H

    E. Ferrara, H. Chang, E. Chen, G. Muric, J. Patel, Characterizing social media manipulation in the 2020 us presidential election, First Monday (2020)

    2020

  18. [26]

    M. M. Akhtar, R. Masood, M. Ikram, S. S. Kanhere, False information, bots and malicious campaigns: Demystifying elements of social media manipulations, arXiv preprint arXiv:2308.12497 (2023)

    2023 arXiv

  19. [27]

    De Angelis, F

    L. De Angelis, F. Baglivo, G. Arzilli, G. P. Privitera, P. Ferragina, A. E. Tozzi, C. Rizzo, Chatgpt and the rise of large language models: the new ai-driven infodemic threat in public health, Frontiers in Public Health 11 (2023) 1166120

    2023

  20. [28]

    C. Chen, K. Shu, Can llm-generated misinformation be detected?, arXiv preprint arXiv:2309.13788 (2023)

    2023 arXiv

  21. [29]

    S. S. Mullappilly, A. Shaker, O. Thawakar, H. Cholakkal, R. M. Anwer, S. Khan, F. S. Khan, Ara- bic mini-climategpt: A climate change and sustainability tailored arabic llm, arXiv preprint arXiv:2312.09366 (2023)

    2023 arXiv

  22. [30]

    R. Vaid, K. Pant, M. Shrivastava, Towards fine-grained classification of climate change related social media text, in: S. Louvan, A. Madotto, B. Madureira (Eds.), Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop...

    2022 doi

  23. [31]

    T. Laud, D. Spokoyny, T. Corringham, T. Berg-Kirkpatrick, Climabench: A benchmark dataset for climate change text understanding in english, arXiv e-prints (2023) arXiv–2301

    2023

  24. [32]

    Pirozelli, M

    P. Pirozelli, M. M. José, I. Silveira, F. Nakasato, S. M. Peres, A. A. F. Brandão, A. H. R. Costa, F. G. Cozman, Benchmarks for pirá 2.0, a reading comprehension dataset about the ocean, the brazilian coast, and climate change, 2023. arXiv:2309.10945

    2023 arXiv

  25. [33]

    Liang, R

    P. Liang, R. Bommasani, T. Lee, D. Tsipras, D. Soylu, M. Yasunaga, Y. Zhang, D. Narayanan, Y. Wu, A. Kumar, et al., Holistic evaluation of language models, arXiv preprint arXiv:2211.09110 (2022)

    2022 arXiv

  26. [34]

    Stiennon, L

    N. Stiennon, L. Ouyang, J. Wu, D. Ziegler, R. Lowe, C. Voss, A. Radford, D. Amodei, P. F. Christiano, Learning to summarize with human feedback, Advances in Neural Information Processing Systems 33 (2020) 3008–3021

    2020

  27. [35]

    W. Zhou, K. Xu, Learning to compare for better training and evaluation of open domain natural language generation models, in: Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 2020, pp. 9717–9724

    2020

  28. [36]

    Ouyang, J

    L. Ouyang, J. Wu, X. Jiang, D. Almeida, C. Wainwright, P. Mishkin, C. Zhang, S. Agarwal, K. Slama, A. Ray, et al., Training language models to follow instructions with human feedback, Advances in neural information processing systems 35 (2022) 27730–27744

    2022

  29. [37]

    P. F. Christiano, J. Leike, T. Brown, M. Martic, S. Legg, D. Amodei, Deep reinforcement learning from human preferences, Advances in neural information processing systems 30 (2017)

    2017

  30. [38]

    URL: https://www.mediabiasfactcheck.com, ac- cessed: 2024-09-09

    MediaBiasFactCheck, Mediabiasfactcheck, 2024. URL: https://www.mediabiasfactcheck.com, ac- cessed: 2024-09-09

    2024

  31. [39]

    URL: https://www.newsguardtech.com/solutions/newsguard/, ac- cessed: 2024-09-09

    NewsGuard, Newsguard, 2024. URL: https://www.newsguardtech.com/solutions/newsguard/, ac- cessed: 2024-09-09

    2024

  32. [40]

    H. Zhu, P. Tiwari, Climate change from large language models, arXiv preprint arXiv:2312.11985 (2023)

    2023 arXiv

  33. [41]

    Kraus, J

    M. Kraus, J. A. Bingler, M. Leippold, T. Schimanski, C. C. Senni, D. Stammbach, S. A. Vaghefi, N. We- bersinke, Enhancing large language models with climate resources, arXiv preprint arXiv:2304.00116 (2023)

    2023 arXiv

  34. [42]

    Achiam, S

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat, et al., Gpt-4 technical report, arXiv preprint arXiv:2303.08774 (2023)

    2023 arXiv

  35. [43]

    Dubey, A

    A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan, et al., The llama 3 herd of models, arXiv preprint arXiv:2407.21783 (2024)

    2024 arXiv

  36. [44]

    Rajpurkar, R

    P. Rajpurkar, R. Jia, P. Liang, Know what you don’t know: Unanswerable questions for squad, arXiv preprint arXiv:1806.03822 (2018)

    2018 arXiv

  37. [45]

    Bavaresco, R

    A. Bavaresco, R. Bernardi, L. Bertolazzi, D. Elliott, R. Fernández, A. Gatt, E. Ghaleb, M. Giulianelli, M. Hanna, A. Koller, et al., Llms instead of human judges? a large scale empirical study across 20 nlp evaluation tasks, arXiv preprint arXiv:2406.18403 (2024)

    2024 arXiv

  38. [46]

    Vidgen, A

    B. Vidgen, A. Agrawal, A. M. Ahmed, V. Akinwande, N. Al-Nuaimi, N. Alfaraj, E. Alhajjar, L. Aroyo, T. Bavalatti, B. Blili-Hamelin, et al., Introducing v0. 5 of the ai safety benchmark from mlcommons, arXiv preprint arXiv:2404.12241 (2024)

    2024 arXiv

  39. [47]

    S. Matz, J. Teeny, S. S. Vaid, H. Peters, G. Harari, M. Cerf, The potential of generative ai for personalized persuasion at scale, Scientific Reports 14 (2024) 4692

    2024

  40. [48]

    code ":

    C. Rojas, F. Algra-Maschio, M. Andrejevic, T. Coan, J. Cook, Y.-F. Li, Augmented cards: A machine learning approach to identifying triggers of climate change misinformation on twitter, arXiv preprint arXiv:2404.15673 (2024). A. Appendix A.1. Climate Change Keywords A full list...

    2024 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.