Pith. sign in

REVIEW 4 major objections 8 minor 45 references

Evaluating LLMs Capabilities Towards Understanding Social Dynamics

T0 review · 4 major / 8 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 LLMs can follow reply chains but not social media slang.

desk verdict The paper's negative finding on LLM social understanding is worth attention, but the one big positive result—PEFT directionality—is not trustworthy as reported because the evaluation set is not separated from the fine-tuning corpus. read the letter →

arxiv 2411.13008 v1 pith:QAIW3I23 submitted 2024-11-20 cs.LG cs.AI

Anique Tahir , Lu Cheng , Manuel Sandoval , Yasin N. Silva , Deborah L. Hall , Huan Liu This is my paper · ORCID
classification cs.LGcs.AI
keywords largelanguagemodelssocialmediadynamicscyberbullyingdetectionanti-bullyingparaphraseevaluationdirectionalityunderstandingparameter-efficientfine-tuningsemantic
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 asks whether generative LLMs can understand the dynamics of informal social media conversations, using cyberbullying and anti-bullying as test cases. It argues that the bottleneck is semantic understanding: even fine-tuned models can learn to track who a comment replies to, but they paraphrase poorly and classify bullying and anti-bullying comments close to chance. The paper's evidence combines paraphrase similarity metrics with edit distance, constrained prompting, and two-phase parameter-efficient fine-tuning. A sympathetic reader would take the finding as a warning that deploying LLMs as social moderators or analysts requires better informal-language comprehension, not just more prompting.

What carries the argument

The evaluation machinery has three parts. For language understanding, the paper pairs BERT-based semantic similarity with the Levenshtein ratio between the generated paraphrase and the original comment; the two together expose verbatim copying that n-gram metrics alone would hide. For directionality, it uses a two-phase PEFT process with LoRA: first WikiTableQuestions data to infuse structural reasoning, then social media data to add social understanding. For classification, it uses constrained generation with a prompt template that asks for a summary as a chain-of-thought step, then binary labels and explanations. This combination lets the paper separate understanding the words from following the conversational structure.

What would settle it

Collect human reference paraphrases for the same Instagram comments and have human raters judge whether model rewrites preserve meaning; if fine-tuned models match those references well, or if cyberbullying and anti-bullying accuracy rises above chance after training on a large informal-language corpus, the paper's bottleneck claim would need revision.

Watch

Extended reading notes

Core claim

On its own terms, the central claim is that the weakest link in using LLMs for social behavior analysis is the lack of semantic understanding of the informal language used in social media. The paper reaches this by comparing GPT-2, Llama-2 7B/13B, and ChatGPT on paraphrasing, directionality, and bullying/anti-bullying detection. It finds that ChatGPT balances semantic similarity and edit distance well in paraphrase tasks, while Llama models often repeat input verbatim and GPT-2 produces gibberish. After parameter-efficient fine-tuning, Llama models significantly improve at identifying which post a comment targets on 4chan, but on Instagram cyberbullying and anti-bullying classification all model variations perform near random chance. The paper concludes that structural and directional understanding can be taught, while social semantics remain the limiting factor.

Load-bearing premise

The argument leans on treating BERT semantic similarity and Levenshtein ratio against the original comment as valid evidence of whether a model understands informal social language, with no human-written reference paraphrases to compare against.

Editorial extensions

If this is right

  • Fine-tuning and exemplar prompting can improve structural tasks such as directionality, but they do not by themselves fix semantic understanding of informal language.
  • Paraphrase evaluations that rely only on BLEU, ROUGE, or Jaccard scores can overstate understanding; the paper shows that edit distance must be paired with semantic similarity to detect verbatim copying.
  • LLM-based cyberbullying and anti-bullying classifiers operating on short informal comments should not be trusted at their current accuracy levels.
  • Closed models' safety filters block this kind of social-media evaluation, so progress depends on open models and better benchmark datasets.
  • A SQuAD-like social language comprehension dataset is needed to measure and drive improvement in informal-language understanding.

Reading between the lines

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

  • If the bottleneck is semantics rather than structure, then injecting more structural reasoning data will not help social classification; the next gains should come from training on informal text with human reference paraphrases.
  • The near-chance classification combined with high verbatim paraphrase rates suggests models may rely on surface lexical cues; a testable extension is to see whether slang-heavy or obfuscated toxic comments reduce accuracy further.
  • Because ChatGPT was excluded from classification due to safety filters, the comparison is incomplete; evaluating an open model under the same constrained-generation setup could separate model capability from policy restrictions.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 8 minor

Summary. The paper evaluates the ability of generative LLMs (GPT-2, Llama-2 7B/13B, and ChatGPT) to understand social media dynamics, using cyberbullying and anti-bullying as case studies. It considers three tasks: paraphrasing informal social media comments (Section 4), identifying directionality of replies in 4chan threads (Section 5), and classifying Instagram comments as cyberbullying or anti-bullying (Section 6). The authors use prompt engineering, exemplars, and two-phase LoRA-based PEFT. They report that fine-tuned models show promising performance on directionality (Table 2), but that all models struggle with semantic understanding and perform near chance on cyberbullying/anti-bullying classification (Tables 1 and 3). The central conclusion is that the weakest link is the lack of semantic understanding of informal social media language.

Significance. If the findings are valid, this paper provides a useful empirical benchmark for LLM capabilities in a socially important setting, distinguishing between structural (directionality) and semantic understanding. The study covers multiple models, prompt variants, and fine-tuning strategies, and it makes a credible case that existing LLMs are not reliable for semantic analysis of informal social media discourse. The paper also highlights a practical problem: closed models like ChatGPT cannot be used for constrained generation on toxic content, limiting their applicability. However, the current evidence is weakened by several experimental gaps, so the contribution is promising but not yet fully established.

major comments (4)
  1. [§5.2 and §6] The directionality results in Table 2 are not interpretable without a train/test split. Section 5.2 states that the 4chan corpus is used for the directionality analysis, and Section 6 later identifies the same 4chan corpus as the basis for 'Phase 2' of PEFT. If the evaluation in Table 2 is performed on threads that were also used for fine-tuning, the reported gains (7B target accuracy 0.082→0.865; 13B 0.159→0.971) may reflect memorization rather than improved structural understanding. Please report the split between training and evaluation threads, the number of threads and target/reply instances, and confirm that no evaluation thread appears in the PEFT training data. If no split exists, the experiment must be re-run on held-out threads before the directionality claim can be accepted.
  2. [§4.1, §4.2, Table 1] The paraphrase-quality evaluation lacks a human reference. All similarity metrics (BLEU, ROUGE, Jaccard, BERT semantic similarity) are computed between the model generation and the original comment, rather than between the generation and human reference paraphrases. As the authors note, verbatim reproduction can therefore inflate all of these scores, and the Levenshtein ratio is only a partial remedy. In particular, the claim that ChatGPT 'shows a robust understanding of social context' (Section 4.2) is not warranted without human-written reference paraphrases or a human evaluation of the generated paraphrases. Please add a human reference set or at least a human rating study.
  3. [§6, Table 3] The cyberbullying/anti-bullying experiment is severely under-powered and lacks statistical support. Only 100 labeled sessions are used, and the paper does not report the number of comments per class, confidence intervals, or any significance test. Since all accuracy values in Table 3 are within 0.513 of 0.5, the conclusion that the models 'perform close to random chance' is not statistically grounded. Please report the number of test instances, class distribution, and error bars or a binomial test, and ideally increase the evaluation set.
  4. [§3.3, §5.2] The fine-tuning configuration is not described in sufficient detail to assess or reproduce the PEFT results. Section 3.3 mentions LoRA and an Alpaca mix probability, but does not give the LoRA rank r, learning rate, batch size, number of epochs, or the exact mix probability. Section 5.2 does not state how many 4chan threads were used for Phase 2, nor how the structural Phase 1 data (WikiTableQuestions) was combined. Please provide these hyperparameters and data statistics.
minor comments (8)
  1. [Abstract] The phrase 'a understanding' should be 'an understanding'.
  2. [§4.1] The metric is 'ROUGE', not 'ROGUE', throughout the section and tables.
  3. [§4.2] There is a typo: 'Levenshein' should be 'Levenshtein'.
  4. [§4.2] The word 'dateset' should be 'dataset'.
  5. [§3.3] The Alpaca instruction data is cited to reference [36], which is the LLaMA paper; please cite the actual Alpaca dataset (Taori et al., 2023).
  6. [Fig. 2] The violin plot axes are not fully described; please clarify what quantity is on each axis and how the distribution is computed.
  7. [§6] Please clarify whether the '100 labeled sessions' means 100 Instagram posts with all their comments, and report the total number of comments used for the binary classification.
  8. [§5.2] The term 'JORA' is used without definition; while reference [34] is given, a brief description of this fine-tuning library would improve readability.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: the paper is an empirical evaluation whose semantic-understanding finding is independently corroborated; the 4chan train/test split gap is a leakage risk, not a demonstrated circularity.

full rationale

This paper is an empirical evaluation rather than a formal derivation, and I find no step in which a predicted quantity is equivalent by construction to a fitted input or to a self-citation. The central semantic-understanding claim (Section 7) is supported by two complementary measures in Section 4.1 -- BERT semantic similarity and Levenshtein ratio -- which jointly distinguish genuine paraphrase from verbatim repetition, and it is independently corroborated by the near-random cyberbullying/anti-bullying classification results in Table 3. Section 6 explicitly states that a separate Instagram dataset was used to avoid data leakage for that evaluation. The directionality result in Table 2 is based on 4chan data, and Section 6 identifies 4chan as the Phase 2 PEFT corpus; the paper does not report a train/test split for the directionality evaluation, which is a legitimate correctness and memorization risk. However, the paper never states that the directionality evaluation examples were drawn from the PEFT training set, so the reduction is not exhibited by the text itself and cannot be flagged as circular under the requirement to quote a specific equivalence. The self-citations present -- JORA [34] and the Instagram dataset [16] -- are tooling and data resources, not load-bearing arguments whose conclusions reduce to the authors' prior claims. Score 1 reflects the minor self-citations and the unresolved leakage concern, not any demonstrated definitional circularity.

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

The paper's claims rest on four domain assumptions about measurement validity and data reliability, plus several unreported fine-tuning hyperparameters. No mathematical derivation or fitted physical constants are involved.

free parameters (4)
  • LoRA rank r = not reported
    Used in both PEFT phases; rank determines adapter size, but the paper never reports the value, so the directionality and classification results cannot be reproduced.
  • PEFT learning rate and epochs = not reported
    Fine-tuning details for Phase 1 (WikiTableQuestions) and Phase 2 (social media) are not given; these choices affect the reported accuracy.
  • Alpaca mix probability = not reported
    Section 3.3 says task data and Alpaca instructions are combined 'with a certain probability' without specifying it.
  • Constrained generation thresholds = multiple (e.g., len(SUMMARY) < 100)
    The prompt constraints in Fig. 1 impose length and format limits that shape model outputs, but their values appear ad hoc and are not justified.
assumptions (4)
  • domain assumption Paraphrase similarity metrics (BLEU, ROUGE, Jaccard, BERT semantic similarity, Levenshtein ratio) measure language understanding.
    Section 4.1 defines the metrics; Section 4.2 uses them to conclude LLMs lack social-context understanding. No human paraphrase references are used.
  • domain assumption 4chan quote/reply tags are reliable ground truth for directionality.
    Section 5.2 uses the 4chan corpus to train and evaluate directionality, assuming the tags correctly indicate the target of each reply.
  • domain assumption The 100 labeled Instagram sessions are accurate and representative for cyberbullying and anti-bullying classification.
    Section 6 uses this dataset for the classification evaluation; the paper does not describe label quality, inter-annotator agreement, or how the 100 sessions were sampled.
  • domain assumption Fine-tuning on WikiTableQuestions imparts structural reasoning that transfers to social media directionality.
    Section 5.1 introduces the surrogate dataset and assumes the learned table reasoning generalizes to comment-level conversation structure.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluating LLMs Capabilities Towards Understanding Social Dynamics." pith.science (2026). https://pith.science/paper/QAIW3I23

@misc{pith2026241113008,
  author       = {Pith},
  title        = {Pith review of: Evaluating LLMs Capabilities Towards Understanding Social Dynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QAIW3I23}},
  note         = {Machine review of arXiv:2411.13008}
}
read the original abstract

Social media discourse involves people from different backgrounds, beliefs, and motives. Thus, often such discourse can devolve into toxic interactions. Generative Models, such as Llama and ChatGPT, have recently exploded in popularity due to their capabilities in zero-shot question-answering. Because these models are increasingly being used to ask questions of social significance, a crucial research question is whether they can understand social media dynamics. This work provides a critical analysis regarding generative LLM's ability to understand language and dynamics in social contexts, particularly considering cyberbullying and anti-cyberbullying (posts aimed at reducing cyberbullying) interactions. Specifically, we compare and contrast the capabilities of different large language models (LLMs) to understand three key aspects of social dynamics: language, directionality, and the occurrence of bullying/anti-bullying messages. We found that while fine-tuned LLMs exhibit promising results in some social media understanding tasks (understanding directionality), they presented mixed results in others (proper paraphrasing and bullying/anti-bullying detection). We also found that fine-tuning and prompt engineering mechanisms can have positive effects in some tasks. We believe that a understanding of LLM's capabilities is crucial to design future models that can be effectively used in social applications.

Figures

Figures reproduced from arXiv: 2411.13008 by the authors.

Figure 1
Figure 1. We use constrained generation for our analysis. Constrained prompt generation allows [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Semantic similarity against edit distance for paraphrases. ChatGPT shows a promising [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. We divide our fine-tuning analysis in two phases. The first phase adds structural under [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 18 canonical work pages

  1. [1]

    arXiv preprint arXiv:1908.10063 (2019)

    Araci, D.: Finbert: Financial sentiment analysis with pre-trained language models. arXiv preprint arXiv:1908.10063 (2019)

    arXiv 2019

  2. [2]

    Atske, S.: The state of online harassment (2021), https://www.pewresearch.org/ internet/2021/01/13/the-state-of-online-harassment/

    work page 2021

  3. [3]

    In: Proceedings of the 13th international workshop on semantic evalua- tion

    Basile, V ., Bosco, C., Fersini, E., Nozza, D., Patti, V ., Pardo, F.M.R., Rosso, P., Sanguinetti, M.: Semeval-2019 task 5: Multilingual detection of hate speech against immigrants and women in twitter. In: Proceedings of the 13th international workshop on semantic evalua- tion. pp. 54–63 (2019)

    work page 2019

  4. [4]

    https://doi.org/ 10.48550/arXiv.2303.12712, http://arxiv.org/abs/2303.12712 14 A

    Bubeck, S., Chandrasekaran, V ., Eldan, R., Gehrke, J., Horvitz, E., Kamar, E., Lee, P., Lee, Y .T., Li, Y ., Lundberg, S., Nori, H., Palangi, H., Ribeiro, M.T., Zhang, Y .: Sparks of arti- ficial general intelligence: Early experiments with GPT-4 (2023). https://doi.org/ 10.48550/arXiv.2303.12712, http://arxiv.org/abs/2303.12712 14 A. Tahir et al

    arXiv doi:10.48550/arxiv.2303.12712 2023

  5. [5]

    In: Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

    Calderón, F.H., Cheng, L.K., Lin, M.J., Huang, Y .H., Chen, Y .S.: Content-based echo chamber detection on social media platforms. In: Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining. p. 597–600. ASONAM ’19, Association for Computing Machinery, New York, NY , USA (2020). https://doi.org/10.1145/...

    work page arXiv 2020

  6. [6]

    In: Proceedings of the 13th international workshop on semantic evaluation

    Chatterjee, A., Narahari, K.N., Joshi, M., Agrawal, P.: Semeval-2019 task 3: Emocontext contextual emotion detection in text. In: Proceedings of the 13th international workshop on semantic evaluation. pp. 39–48 (2019)

    work page 2019

  7. [7]

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

    arXiv 2023

  8. [8]

    In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining

    Cheng, L., Guo, R., Shu, K., Liu, H.: Causal understanding of fake news dissemination on social media. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. p. 148–157. KDD ’21, Association for Computing Machinery, New York, NY , USA (2021). https://doi.org/10.1145/3447548.3467321, https://doi.org/10.1145/3447548.3467321

    arXiv 2021

Show all 45 references
  1. [9]

    arXiv preprint arXiv:1905.12516 (2019)

    Davidson, T., Bhattacharya, D., Weber, I.: Racial bias in hate speech and abusive language detection datasets. arXiv preprint arXiv:1905.12516 (2019)

    2019 arXiv

  2. [10]

    arXiv preprint arXiv:1810.04805 (2018)

    Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)

    2018 arXiv

  3. [11]

    In: Gervasi, O., Murgante, B., Misra, S., Rocha, A.M.A.C., Garau, C

    Duzen, Z., Riveni, M., Aktas, M.S.: Misinformation detection in social networks: A system- atic literature review. In: Gervasi, O., Murgante, B., Misra, S., Rocha, A.M.A.C., Garau, C. (eds.) Computational Science and Its Applications – ICCSA 2022 Workshops. pp. 57–74. Springer...

    2022

  4. [12]

    arXiv preprint arXiv:1708.00524 (2017)

    Felbo, B., Mislove, A., Søgaard, A., Rahwan, I., Lehmann, S.: Using millions of emoji oc- currences to learn any-domain representations for detecting sentiment, emotion and sarcasm. arXiv preprint arXiv:1708.00524 (2017)

    2017 arXiv

  5. [13]

    arXiv preprint arXiv:1901.08458 (2019)

    Gaind, B., Syal, V ., Padgalwar, S.: Emotion detection and analysis on social media. arXiv preprint arXiv:1901.08458 (2019)

    2019 arXiv

  6. [14]

    In: Proceedings of the 2018 World Wide Web Con- ference

    Gillani, N., Yuan, A., Saveski, M., V osoughi, S., Roy, D.: Me, my echo chamber, and i: In- trospection on social media polarization. In: Proceedings of the 2018 World Wide Web Con- ference. p. 823–831. WWW ’18, International World Wide Web Conferences Steering Com- mittee, Re...

    2018

  7. [15]

    org/10.1016/j.wneu.2023.08.042, https://www.sciencedirect.com/ science/article/pii/S1878875023011440

    Guerra, G.A., Hofmann, H., Sobhani, S., Hofmann, G., Gomez, D., Soroudi, D., Hopkins, B.S., Dallas, J., Pangal, D.J., Cheok, S., Nguyen, V .N., Mack, W.J., Zada, G.: GPT-4 artificial intelligence model outperforms ChatGPT, medical students, and neurosurgery residents on neuros...

    2023 doi

  8. [16]

    In: Proceedings of the International AAAI Conference on Web and Social Media

    Hamlett, M., Powell, G., Silva, Y .N., Hall, D.: A labeled dataset for investigating cyberbul- lying content patterns in instagram. In: Proceedings of the International AAAI Conference on Web and Social Media. vol. 16, pp. 1251–1258 (2022)

    2022

  9. [17]

    https://doi.org/10.48550/arXiv.2303.16854, http:// arxiv.org/abs/2303.16854

    He, X., Lin, Z., Gong, Y ., Jin, A.L., Zhang, H., Lin, C., Jiao, J., Yiu, S.M., Duan, N., Chen, W.: AnnoLLM: Making large language models to be better crowdsourced an- notators (2023). https://doi.org/10.48550/arXiv.2303.16854, http:// arxiv.org/abs/2303.16854

    2023 doi arXiv

  10. [18]

    In: Social Informatics: 7th International Conference, SocInfo 2015, Beijing, China, December 9-12, 2015, Proceedings

    Hosseinmardi, H., Mattson, S.A., Ibn Rafiq, R., Han, R., Lv, Q., Mishra, S.: Analyzing la- beled cyberbullying incidents on the instagram social network. In: Social Informatics: 7th International Conference, SocInfo 2015, Beijing, China, December 9-12, 2015, Proceedings

    2015

  11. [19]

    pp. 49–66. Springer (2015) Evaluating LLMs Capabilities Towards Understanding Social Dynamics 15

    2015

  12. [20]

    arXiv preprint arXiv:2106.09685 (2021)

    Hu, E.J., Shen, Y ., Wallis, P., Allen-Zhu, Z., Li, Y ., Wang, S., Wang, L., Chen, W.: Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685 (2021)

    2021 arXiv

  13. [21]

    arXiv preprint arXiv:2305.00050 (2023)

    Kıcıman, E., Ness, R., Sharma, A., Tan, C.: Causal reasoning and large language models: Opening a new frontier for causality. arXiv preprint arXiv:2305.00050 (2023)

    2023 arXiv

  14. [22]

    Com- puters in Human Behavior 130, 107200 (May 2022)

    Lan, M., Law, N., Pan, Q.: Effectiveness of anti-cyberbullying educational pro- grams: A socio-ecologically grounded systematic review and meta-analysis. Com- puters in Human Behavior 130, 107200 (May 2022). https://doi.org/10. 1016/j.chb.2022.107200, https://www.sciencedirect...

    2022

  15. [23]

    Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., Neubig, G.: Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing (2021)

    2021

  16. [24]

    https://doi.org/ 10.1609/aaai.v32i1.11268, https://ojs.aaai.org/index.php/AAAI/ article/view/11268, number: 1

    Liu, Y ., Wu, Y .F.: Early detection of fake news on social media through propagation path classification with recurrent and convolutional networks32(1). https://doi.org/ 10.1609/aaai.v32i1.11268, https://ojs.aaai.org/index.php/AAAI/ article/view/11268, number: 1

    doi

  17. [25]

    Journal of Experimental & Theoretical Artificial Intelligence 30(2), 187–202 (2018)

    Malmasi, S., Zampieri, M.: Challenges in discriminating profanity from hate speech. Journal of Experimental & Theoretical Artificial Intelligence 30(2), 187–202 (2018)

    2018

  18. [26]

    arXiv preprint arXiv:2110.15943 (2021)

    Min, S., Lewis, M., Zettlemoyer, L., Hajishirzi, H.: Metaicl: Learning to learn in context. arXiv preprint arXiv:2110.15943 (2021)

    2021 arXiv

  19. [27]

    In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management

    Minici, M., Cinus, F., Monti, C., Bonchi, F., Manco, G.: Cascade-based echo chamber detection. In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management. p. 1511–1520. CIKM ’22, Association for Computing Machinery, New York, NY , USA (2022)...

    2022

  20. [28]

    arXiv preprint arXiv:2005.10200 (2020)

    Nguyen, D.Q., Vu, T., Nguyen, A.T.: Bertweet: A pre-trained language model for english tweets. arXiv preprint arXiv:2005.10200 (2020)

    2020 arXiv

  21. [29]

    arXiv preprint arXiv:2305.01651 (2023)

    Onoe, Y ., Zhang, M.J., Padmanabhan, S., Durrett, G., Choi, E.: Can lms learn new en- tities from descriptions? challenges in propagating injected knowledge. arXiv preprint arXiv:2305.01651 (2023)

    2023 arXiv

  22. [30]

    In: Proceedings of the international AAAI conference on web and social media

    Papasavva, A., Zannettou, S., De Cristofaro, E., Stringhini, G., Blackburn, J.: Raiders of the lost kek: 3.5 years of augmented 4chan posts from the politically incorrect board. In: Proceedings of the international AAAI conference on web and social media. vol. 14, pp. 885–894 (2020)

    2020

  23. [31]

    arXiv preprint arXiv:1508.00305 (2015)

    Pasupat, P., Liang, P.: Compositional semantic parsing on semi-structured tables. arXiv preprint arXiv:1508.00305 (2015)

    2015 arXiv

  24. [32]

    CoRR abs/1806.03822 (2018), http://arxiv.org/abs/1806.03822

    Rajpurkar, P., Jia, R., Liang, P.: Know what you don’t know: Unanswerable questions for squad. CoRR abs/1806.03822 (2018), http://arxiv.org/abs/1806.03822

    2018 arXiv

  25. [33]

    In: Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments

    Risch, J., Stoll, A., Wilms, L., Wiegand, M.: Overview of the germeval 2021 shared task on the identification of toxic, engaging, and fact-claiming comments. In: Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. ...

    2021

  26. [34]

    Computers in Human Behavior 93, 333–345 (2019)

    Rosa, H., Pereira, N., Ribeiro, R., Ferreira, P., Carvalho, J., Oliveira, S., Coheur, L., Paulino, P., Veiga Simão, A., Trancoso, I.: Automatic cyberbullying detection: A systematic re- view. Computers in Human Behavior 93, 333–345 (2019). https://doi.org/https: //doi.org/10.1...

    2019 doi

  27. [35]

    arXiv preprint arXiv:2403.11366 (2024)

    Tahir, A., Cheng, L., Liu, H.: Jora: Jax tensor-parallel lora library for retrieval augmented fine-tuning. arXiv preprint arXiv:2403.11366 (2024)

    2024 arXiv

  28. [36]

    2102.02503, http://arxiv.org/abs/2102.02503 16 A

    Tamkin, A., Brundage, M., Clark, J., Ganguli, D.: Understanding the capabilities, limitations, and societal impact of large language models.https://doi.org/10.48550/arXiv. 2102.02503, http://arxiv.org/abs/2102.02503 16 A. Tahir et al

    doi arXiv

  29. [37]

    https://doi

    Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.A., Lacroix, T., Rozière, B., Goyal, N., Hambro, E., Azhar, F., Rodriguez, A., Joulin, A., Grave, E., Lample, G.: LLaMA: Open and efficient foundation language models (2023). https://doi. org/10.48550/arXiv.2302.13...

    2023 doi arXiv

  30. [38]

    arXiv preprint arXiv:2302.06706 (2023)

    Valmeekam, K., Sreedharan, S., Marquez, M., Olmo, A., Kambhampati, S.: On the planning abilities of large language models (a critical investigation with a proposed benchmark). arXiv preprint arXiv:2302.06706 (2023)

    2023 arXiv

  31. [39]

    Advances in neural information processing systems 30 (2017)

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in neural information processing systems 30 (2017)

    2017

  32. [40]

    New Media & Society23(6), 1379–1397 (2021)

    Wang, S.: Standing up or standing by: Bystander intervention in cyberbullying on social media. New Media & Society23(6), 1379–1397 (2021). https://doi.org/10.1177/ 1461444820902541, https://doi.org/10.1177/1461444820902541

    2021 doi

  33. [41]

    arXiv preprint arXiv:2206.07682 (2022)

    Wei, J., Tay, Y ., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., et al.: Emergent abilities of large language models. arXiv preprint arXiv:2206.07682 (2022)

    2022 arXiv

  34. [42]

    Advances in Neural Information Processing Systems 35, 24824–24837 (2022)

    Wei, J., Wang, X., Schuurmans, D., Bosma, M., Xia, F., Chi, E., Le, Q.V ., Zhou, D., et al.: Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems 35, 24824–24837 (2022)

    2022

  35. [43]

    arXiv preprint arXiv:2004.11493 (2020)

    Wiedemann, G., Yimam, S.M., Biemann, C.: Uhh-lt at semeval-2020 task 12: Fine- tuning of pre-trained transformer networks for offensive language detection. arXiv preprint arXiv:2004.11493 (2020)

    2020 arXiv

  36. [44]

    https://doi.org/10.48550/arXiv.2309

    Ye, H., Liu, T., Zhang, A., Hua, W., Jia, W.: Cognitive mirage: A review of hallu- cinations in large language models. https://doi.org/10.48550/arXiv.2309. 06794, http://arxiv.org/abs/2309.06794

    doi arXiv

  37. [45]

    Expert Systems with Applications 170, 114547 (2021)

    Zotova, E., Agerri, R., Rigau, G.: Semi-automatic generation of multilingual datasets for stance detection in twitter. Expert Systems with Applications 170, 114547 (2021)

    2021

Pith tools

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