Pith. sign in

REVIEW 4 major objections 5 minor 58 references

On the Adaptive Psychological Persuasion of Large Language Models

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a preference-optimization training loop over strategy-specific persuasive responses lets LLMs autonomously choose contextually effective psychological tactics, raising success rates on held-out counterfactuals after…

desk verdict Worth reading and worth reviewing: real empirical work on adaptive strategy selection, but the DPO equation as written isn't the DPO objective, so the released code or a corrected formal description is needed before the central claim is fully supported. read the letter →

arxiv 2506.06800 v1 pith:R4AJ5DRG submitted 2025-06-07 cs.CL

Tianjie Ju , Yujia Chen , Hao Fei , Mong-Li Lee , Wynne Hsu , Pengzhou Cheng , Zongru Wu , Zhuosheng Zhang
show 1 more author
Gongshen Liu
This is my paper
classification cs.CL
keywords psychologicalpersuasionlargelanguagemodelsdirectpreferenceoptimizationcounterfactualbeliefeditingsuccessrateepistemicresistancestrategyselection
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 large language models can not only produce persuasive rhetoric but also choose which psychological tactic fits a given counterfactual claim. It first shows that unguided persuader LLMs lean on repetitive, weak strategies, and that explicitly prompting one of eleven psychological strategies helps but that no single strategy wins across contexts. The central proposal is an adaptive training loop: for each counterfactual, the model's attempts under different strategies are scored by whether the listener accepts the alternate object, and direct preference optimization (DPO) is used to make the persuader favor strategies that worked. With 3,000 training examples, the fine-tuned models outperform their originals under almost every strategy instruction and shift toward a more diverse, context-dependent mix of strategies, while MMLU scores stay essentially flat. The paper concludes that adaptive strategy selection is a viable route to stronger and more flexible machine persuasion.

What carries the argument

The load-bearing mechanism is an adaptive preference-optimization loop built on direct preference optimization (DPO), a training objective that increases the probability of preferred responses and decreases dispreferred ones. Here the 'preferences' are not human labels but outcomes: for each counterfactual $(s, r, o^*)$, the persuader generates messages under several strategies, and the listener's acceptance or rejection turns each message into a win or loss. Successful and failed strategy-conditioned responses become the positive and negative examples in the DPO objective, with five such pairs sampled per training instance, and instances with no mix of success and failure are discarded. The authors claim this teaches the persuader to autonomously favor strategies that work in a given semantic context. The eleven-strategy toolbox (fluency, framing, authority, scarcity, repetition, conformity, flattery, information isolation, argument quantity, argument quality, affective forecasting) supplies the action space the model learns to choose from.

What would settle it

Train the same three models with a corrected DPO objective that includes the persuasive message and a frozen reference model; if the Table 3 improvements over the unguided baseline disappear, the claimed mechanism of learning strategy preferences from persuasion outcomes is not what drives the gains. A second check is to shuffle the success/failure labels and see whether persuasion success still rises, which would indicate the effect is not preference-driven.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that persuasion success in LLMs is a context-sensitive function of rhetorical strategy, and that this function can be learned. The authors define persuasion as updating a listener model's belief from a true object $o$ to a counterfactual object $o^*$ for a factual triple $(s, r, o)$. They assemble a toolbox of eleven psychologically grounded strategies and prompt LLMs to adopt each one; across four LLMs and four semantic domains (person, geography, culture, life), they find large but inconsistent gains, so no static strategy dominates. The paper then fine-tunes three open-source LLMs with DPO, constructing preference pairs from strategy-specific responses that did or did not win over the listener. After one LoRA epoch on 3,000 samples, the trained persuaders beat their untrained versions under nearly all strategy instructions, pick a broader and more effective strategy mix when left unguided, and retain their MMLU performance.

Load-bearing premise

The load-bearing premise is that the printed DPO objective is a valid preference optimization over persuasive strategies; the equation as written compares probabilities of the counterfactual object given only subject and relation, omitting the persuasive message and any reference model, so the training signal it describes may not be the strategy-selection signal the paper relies on.

Editorial extensions

If this is right

  • If the central claim is correct, explicit strategy prompting is not the ceiling: a model trained on its own persuasion outcomes can exceed every fixed strategy, including the best static one.
  • Adaptively trained persuaders remain within noise on MMLU, so the gains are not bought by a collapse of general reasoning.
  • Because the training signal comes from in-model interactions, the same pipeline can be applied to any open-weight model without external preference labels.
  • The learned behavior transfers to unguided generation, producing a more diversified set of strategies on held-out counterfactuals.
  • The result suggests that persuasion capability is better modeled as a per-context policy than as a fixed skill.

Reading between the lines

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

  • Beyond the paper: the same outcome-pair loop could be reversed to train listeners to resist persuasion, using failed persuasion attempts as negative examples and yielding an epistemic-resistance policy.
  • Beyond the paper: the measured strategy diversity depends on GPT-4o's classification of which strategy a message uses; a human-annotated or open-label variant could change Table 4.
  • Beyond the paper: if the adaptive mechanism holds up, it is dual-use, since the same 3,000-example recipe could be aimed at misinformation campaigns; safety work should explore contrastive training that keeps listeners rigid.
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 / 5 minor

Summary. The paper studies whether LLMs can autonomously persuade and resist persuasion on counterfactual edits from the COUNTERFACT dataset. It first measures baseline dual capabilities of four LLMs, then introduces eleven psychological persuasion strategies and shows that strategy effectiveness varies by model and by semantic domain, and finally fine-tunes three open-source models with a DPO-style objective on strategy-specific preference pairs to learn adaptive strategy selection. The main claims are that DPO-trained models improve persuasion success rates on held-out data, choose a more diverse set of strategies without explicit instructions, and maintain general capabilities as measured by MMLU. The empirical narrative is clear, but the formal training objective in Eq. (2) is not the standard DPO loss as written, and the headline gains lack confidence intervals or significance tests.

Significance. If validated, the adaptive framework is a useful contribution: it turns persuasion outcomes into preference pairs, reports systematic comparisons across four LLMs and four semantic domains, and includes an MMLU sanity check. The paper also provides a useful negative result that no static psychological strategy dominates across all contexts. The positive claim that 3,000 training examples suffice to improve autonomous strategy selection is plausible, but it currently rests on unvalidated GPT-4o annotations, internally inconsistent percentages in Table 4, and a training-objective equation that does not match standard DPO. These issues need to be resolved before the central empirical claims can be accepted.

major comments (4)
  1. [§3.3, Eq. (2)] The DPO objective as written compares log p_{π+}(o*|s,r) with log p_{π−}(o*|s,r). This is not the direct preference optimization loss: it does not contain the persuader message m, the trained policy's likelihood, or a frozen reference policy, all of which appear in the canonical DPO loss over preferred and dispreferred completions. Since Section 4.4's before/after PSR differences are the paper's central evidence, Eq. (2) must be corrected to the actual objective used in the released code. If the implementation literally matches Eq. (2), then the reported gains cannot be attributed to DPO-based strategy selection.
  2. [§4.4, Table 3] The claim that adaptive training "significantly enhances" success rates is not supported by any confidence intervals, significance tests, or multiple-seed variance. The table contains several sizable decreases (Qwen Authority Effect −7.29, Qwen Argument Quantity −4.95, Falcon Authority Effect −4.74), and many increases are under two percentage points; without error bars it is impossible to tell which differences exceed noise. Please report bootstrap intervals or multiple training runs.
  3. [Table 4] Several percentages in Table 4 are internally inconsistent: LLaMA Authority Effect "After" is reported as 131 (68.26%) and LLaMA Fluency Effect as 59 (30.75%), yet the column totals to roughly 100% only if these are read as 6.83% and 3.07%; Falcon Repetition Effect 5 (0.03%) and Scarcity Effect 4 (0.02%) show similar decimal errors. Since the post-training diversification claim is one of the two main adaptive-training results, these numbers must be corrected and the aggregate statistics recomputed.
  4. [Tables 2 and 4] The semantic-domain partition in Table 2 and the strategy-usage analysis in Table 4 rely entirely on GPT-4o annotations with no validation. The prompt in Appendix C asks for a single strategy label, but no agreement rate, error analysis, or human-annotated sample is reported. Noisy labels could create the appearance of domain-dependent strategy effectiveness or of post-training diversification; please report annotation agreement on a sample and provide representative annotated examples.
minor comments (5)
  1. [§4.3] The terminology is inconsistent: "Flattery Effect" and "Flattery Trap" are used interchangeably, and "Repetitive Effect" appears where "Repetition Effect" is intended; please unify the strategy names throughout.
  2. [§3.3 and §4.1.2] Section 3.3 defines k strategy pairs but Section 4.1.2 states that five pairs are sampled per training instance; please define k and clarify whether pairs are sampled with or without replacement.
  3. [Table 2] The domain counts (844 + 516 + 393 + 161 = 1,914) do not sum to the stated 1,919 evaluation samples; please explain the discrepancy or correct the counts.
  4. [Table 4] The GPT-4o columns list "Before" values but no "After" values; since GPT-4o is not fine-tuned, this should be stated explicitly or the empty columns should be removed.
  5. [Figure 2] The caption says the four LLMs play in a "4 × 4 adversarial game," but Figure 2 appears to plot aggregate persuasion and resistance values; consider clarifying the relationship between Figure 2 and the pairwise heatmap in Figure 3.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the DPO fine-tuning is evaluated on a held-out split and the central claims do not reduce to their inputs by construction.

full rationale

The paper's derivation chain is an empirical pipeline, not a formal derivation that folds its conclusion into its assumptions. Strategy-specific persuasive responses are generated on 3,000 training counterfactuals, their persuasion outcomes are converted into preference pairs, and a DPO-style fine-tune is applied. The main results in Table 3 are then computed on 1,919 held-out samples, and the paper explicitly states: "All reported results are computed on the held-out evaluation set to prevent data leakage." Because the evaluation set is disjoint from the training labels, the reported PSR improvements are not forced by construction; they are an empirical generalization claim of the same type as standard supervised fine-tuning. The training signal and evaluation metric are both derived from persuasion success, but that does not make the result circular once a held-out split is used. The MMLU results in Table 5 provide an external capability check, and the COUNTERFACT dataset is an external benchmark. The self-citations to Ju et al. appear only in related work and limitations, and are not load-bearing for the adaptive framework. The formal weakness in Equation (2) is a correctness and reproducibility concern: as written, the loss omits the generated message, the trained policy, and the reference model, so it is not the canonical DPO objective and the mechanism described in the text is not faithfully represented by the equation. However, this is a validity flaw, not circularity, because it does not make the held-out evaluation equivalent to the training objective. No circular step could be exhibited with a specific reduction of a prediction to its input.

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

The central claim rests mainly on the COUNTERFACT testbed, the strategy taxonomy, GPT-4o annotations, and a binary success signal. No new physical or mathematical entities are introduced. The main hidden load is the untested reliability of the preference labels and of GPT-4o as annotator.

free parameters (4)
  • DPO temperature beta = not reported
    Eq. 2's loss depends on beta, but the paper never states its value or whether it was tuned.
  • LoRA rank / alpha = 128 / 128
    Fine-tuning hyperparameters are fixed without sensitivity analysis; the reported gains could depend on this choice.
  • Learning rate and training length = 1e-4, 1 epoch
    Single configuration reported; no search or ablation and no variance across seeds.
  • Preference pairs per instance k = 5
    The paper randomly samples five success and failure pairs per training instance; the choice of k is not analyzed.
assumptions (4)
  • domain assumption COUNTERFACT is a valid testbed for psychological persuasion evaluation.
    All central numbers come from this dataset, including the 3,000 and 1,919 split; there is no evidence of transfer to other persuasion settings.
  • domain assumption The 11 strategies are distinct and are faithfully elicited by the provided prompts.
    Strategy definitions are taken from psychology literature, but the paper does not verify that generated texts actually instantiate the intended strategy.
  • ad hoc to paper GPT-4o annotations of counterfactual domains and strategy usage are correct.
    No agreement measures or human validation are reported; the strategy distribution table has inconsistent percentages.
  • domain assumption Exact-match listener output o* after a single exchange is a sufficient measure of persuasion success.
    This binary metric ignores partial belief shifts, and the Rephrase Accuracy results show high fragility, yet PSR is treated as the primary signal.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Adaptive Psychological Persuasion of Large Language Models." pith.science (2026). https://pith.science/paper/R4AJ5DRG

@misc{pith2026250606800,
  author       = {Pith},
  title        = {Pith review of: On the Adaptive Psychological Persuasion of Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R4AJ5DRG}},
  note         = {Machine review of arXiv:2506.06800}
}
read the original abstract

Previous work has showcased the intriguing capabilities of Large Language Models (LLMs) in instruction-following and rhetorical fluency. However, systematic exploration of their dual capabilities to autonomously persuade and resist persuasion, particularly in contexts involving psychological rhetoric, remains unexplored. In this paper, we first evaluate four commonly adopted LLMs by tasking them to alternately act as persuaders and listeners in adversarial dialogues. Empirical results show that persuader LLMs predominantly employ repetitive strategies, leading to low success rates. Then we introduce eleven comprehensive psychological persuasion strategies, finding that explicitly instructing LLMs to adopt specific strategies such as Fluency Effect and Repetition Effect significantly improves persuasion success rates. However, no ``one-size-fits-all'' strategy proves universally effective, with performance heavily dependent on contextual counterfactuals. Motivated by these observations, we propose an adaptive framework based on direct preference optimization that trains LLMs to autonomously select optimal strategies by leveraging persuasion results from strategy-specific responses as preference pairs. Experiments on three open-source LLMs confirm that the proposed adaptive psychological persuasion method effectively enables persuader LLMs to select optimal strategies, significantly enhancing their success rates while maintaining general capabilities. Our code is available at https://github.com/KalinaEine/PsychologicalPersuasion.

Figures

Figures reproduced from arXiv: 2506.06800 by the authors.

Figure 1
Figure 1. Introduction of psychological persuasion in LLMs. We introduce 11 psychological strategies [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The dual capabilities of LLMs in autonomously generating psychological persuasion [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Pairwise heatmaps of adversarial dialogues between LLMs acting as persuader and listener. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Persuasion success rates (%) of eleven explicit psychological strategies and the no-strategy [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

58 extracted references · 45 canonical work pages

  1. [1]

    Aggarwal, S

    P. Aggarwal, S. Y . Jun, and J. H. Huh. Scarcity messages.Journal of Advertising, 40(3):19–30, 2011

    work page 2011

  2. [2]

    R. Anil, S. Borgeaud, Y . Wu, J. Alayrac, J. Yu, R. Soricut, J. Schalkwyk, A. M. Dai, A. Hauth, K. Millican, D. Silver, S. Petrov, M. Johnson, I. Antonoglou, J. Schrittwieser, A. Glaese, J. Chen, E. Pitler, T. P. Lillicrap, A. Lazaridou, O. Firat, J. Molloy, M. Isard, P. R. Barham, T. Hennigan, B. Lee, F. Viola, M. Reynolds, Y . Xu, R. Doherty, E. Collins...

    arXiv 2023

  3. [3]

    T. Blass. The milgram paradigm after 35 years: Some things we now know about obedience to authority 1. Journal of applied social psychology , 29(5):955–978, 1999

    work page 1999

  4. [4]

    Bond and P

    R. Bond and P. B. Smith. Culture and conformity: A meta-analysis of studies using asch’s (1952b, 1956) line judgment task. Psychological bulletin, 119(1):111, 1996

    work page 1956

  5. [5]

    S. Chaiken. Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of personality and social psychology , 39(5):752, 1980

    work page 1980

  6. [6]

    N. Chen, Y . Deng, and J. Li. The oscars of AI theater: A survey on role-playing with language models. CoRR, abs/2407.11484, 2024

    arXiv 2024

  7. [7]

    Z. Chen, J. Wu, J. Zhou, B. Wen, G. Bi, G. Jiang, Y . Cao, M. Hu, Y . Lai, Z. Xiong, and M. Huang. Tombench: Benchmarking theory of mind in large language models. In L. Ku, A. Martins, and V . Srikumar, editors,Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers), ACL 2024, Bangkok, Thailand, Aug...

    work page 2024

  8. [8]

    W. Dong, Y . Zhao, Z. Sun, Y . Liu, Z. Peng, J. Zheng, Z. Zhang, Z. Zhang, J. Wu, R. Wang, S. Xu, X. Huang, and X. He. Humanizing llms: A survey of psychological measurements with tools, datasets, and human-agent applications, 2025

    work page 2025

Show all 58 references
  1. [9]

    Dubey, A

    A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan, A. Goyal, A. Hartshorn, A. Yang, A. Mitra, A. Sravankumar, A. Korenev, A. Hinsvark, A. Rao, A. Zhang, A. Rodriguez, A. Gregerson, A. Spataru, B. Rozière, B. Biron, B. Ta...

    2024 arXiv

  2. [10]

    Elazar, N

    Y . Elazar, N. Kassner, S. Ravfogel, A. Ravichander, E. H. Hovy, H. Schütze, and Y . Goldberg. Measuring and improving consistency in pretrained language models. Trans. Assoc. Comput. Linguistics, 9:1012–1031, 2021

    2021

  3. [11]

    Hendrycks, C

    D. Hendrycks, C. Burns, S. Basart, A. Zou, M. Mazeika, D. Song, and J. Steinhardt. Measuring massive multitask language understanding. In 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021 . OpenReview.net, 2021

    2021

  4. [12]

    Hilliard, C

    A. Hilliard, C. Muñoz, Z. Wu, and A. S. Koshiyama. Eliciting personality traits in large language models. CoRR, abs/2402.08341, 2024

    2024 arXiv

  5. [13]

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen. Lora: Low-rank adaptation of large language models. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022 . OpenReview.net, 2022. 11

    2022

  6. [14]

    Jiang, X

    H. Jiang, X. Zhang, X. Cao, C. Breazeal, D. Roy, and J. Kabbara. Personallm: Investigating the ability of large language models to express personality traits. In K. Duh, H. Gómez-Adorno, and S. Bethard, editors, Findings of the Association for Computational Linguistics: NAACL ...

    2024

  7. [15]

    C. Jin, K. Ren, L. Kong, X. Wang, R. Song, and H. Chen. Persuading across diverse domains: a dataset and persuasion large language model. In L. Ku, A. Martins, and V . Srikumar, editors, Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (V...

    2024

  8. [16]

    T. Ju, Z. Shao, B. Wang, Y . Chen, Z. Zhang, H. Fei, M.-L. Lee, W. Hsu, S. Duan, and G. Liu. Probing then editing response personality of large language models, 2025

    2025

  9. [17]

    T. Ju, Y . Wang, X. Ma, P. Cheng, H. Zhao, Y . Wang, L. Liu, J. Xie, Z. Zhang, and G. Liu. Flooding spread of manipulated knowledge in llm-based multi-agent communities. CoRR, abs/2407.07791, 2024

    2024 arXiv

  10. [18]

    seizing

    A. W. Kruglanski and D. M. Webster. Motivated closing of the mind:" seizing" and" freezing.". Psychological review, 103(2):263, 1996

    1996

  11. [19]

    S. Lee, S. Lim, S. Han, G. Oh, H. Chae, J. Chung, M. Kim, B. Kwak, Y . Lee, D. Lee, J. Yeo, and Y . Yu. Do llms have distinct and consistent personality? TRAIT: personality testset designed for llms with psychometrics. CoRR, abs/2406.14703, 2024

    2024 arXiv

  12. [20]

    X. Li, Y . Li, L. Qiu, S. Joty, and L. Bing. Evaluating psychological safety of large language models. In Y . Al-Onaizan, M. Bansal, and Y . Chen, editors,Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024, Miami, FL, USA, Novemb...

    2024

  13. [21]

    S. Mao, X. Wang, M. Wang, Y . Jiang, P. Xie, F. Huang, and N. Zhang. Editing personality for large language models. In D. F. Wong, Z. Wei, and M. Yang, editors, Natural Language Processing and Chinese Computing - 13th National CCF Conference, NLPCC 2024, Hangzhou, China, Novem...

    2024

  14. [22]

    K. Meng, D. Bau, A. Andonian, and Y . Belinkov. Locating and editing factual associations in GPT. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors,Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Process...

    2022

  15. [23]

    GPT-4 technical report

    OpenAI. GPT-4 technical report. CoRR, abs/2303.08774, 2023

    2023 arXiv

  16. [24]

    Rafailov, A

    R. Rafailov, A. Sharma, E. Mitchell, C. D. Manning, S. Ermon, and C. Finn. Direct preference optimization: Your language model is secretly a reward model. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors, Advances in Neural Information Processing...

    2023

  17. [25]

    Rivière, S

    M. Rivière, S. Pathak, P. G. Sessa, C. Hardin, S. Bhupatiraju, L. Hussenot, T. Mesnard, B. Shahri- ari, A. Ramé, J. Ferret, P. Liu, P. Tafti, A. Friesen, M. Casbon, S. Ramos, R. Kumar, C. L. Lan, S. Jerome, A. Tsitsulin, N. Vieillard, P. Stanczyk, S. Girgin, N. Momchev, M. Hof...

    2024 arXiv

  18. [26]

    Sabour, S

    S. Sabour, S. Liu, Z. Zhang, J. M. Liu, J. Zhou, A. S. Sunaryo, T. M. C. Lee, R. Mihalcea, and M. Huang. Emobench: Evaluating the emotional intelligence of large language models. In L. Ku, A. Martins, and V . Srikumar, editors, Proceedings of the 62nd Annual Meeting of the Ass...

    2024

  19. [27]

    Safdari, G

    M. Safdari, G. Serapio-García, C. Crepy, S. Fitz, P. Romero, L. Sun, M. Abdulhai, A. Faust, and M. J. Mataric. Personality traits in large language models. CoRR, abs/2307.00184, 2023

    2023 arXiv

  20. [28]

    Sorokovikova, N

    A. Sorokovikova, N. Fedorova, S. Rezagholi, and I. P. Yamshchikov. Llms simulate big five personality traits: Further evidence. CoRR, abs/2402.01765, 2024

    2024 arXiv

  21. [29]

    Strachan, D

    J. Strachan, D. Albergo, G. Borghini, O. Pansardi, E. Scaliti, S. Gupta, K. Saxena, A. Rufo, S. Panzeri, G. Manzi, M. Graziano, and C. Becchio. Testing theory of mind in large language models and humans. Nature Human Behaviour, 8:1–11, 05 2024

    2024

  22. [30]

    F.-L. Team. The falcon 3 family of open models, December 2024

    2024

  23. [31]

    I. Team. Internlm: A multilingual language model with progressively enhanced capabilities. https://github.com/InternLM/InternLM-techreport, 2023

    2023

  24. [32]

    Q. Tu, S. Fan, Z. Tian, T. Shen, S. Shang, X. Gao, and R. Yan. Charactereval: A chinese benchmark for role-playing conversational agent evaluation. In L. Ku, A. Martins, and V . Sriku- mar, editors, Proceedings of the 62nd Annual Meeting of the Association for Computational Li...

    2024

  25. [33]

    Tversky and D

    A. Tversky and D. Kahneman. The framing of decisions and the psychology of choice. science, 211(4481):453–458, 1981

    1981

  26. [34]

    Unkelbach and R

    C. Unkelbach and R. Greifeneder. A general model of fluency effects in judgment and decision making. In The experience of thinking , pages 11–32. Psychology Press, 2013

    2013

  27. [35]

    R. V onk. Self-serving interpretations of flattery: Why ingratiation works.Journal of personality and social psychology, 82(4):515, 2002

    2002

  28. [36]

    N. Wang, Z. Peng, H. Que, J. Liu, W. Zhou, Y . Wu, H. Guo, R. Gan, Z. Ni, J. Yang, M. Zhang, Z. Zhang, W. Ouyang, K. Xu, W. Huang, J. Fu, and J. Peng. Rolellm: Benchmarking, eliciting, and enhancing role-playing abilities of large language models. In L. Ku, A. Martins, and V ....

    2024

  29. [37]

    P. Wang, N. Zhang, B. Tian, Z. Xi, Y . Yao, Z. Xu, M. Wang, S. Mao, X. Wang, S. Cheng, K. Liu, Y . Ni, G. Zheng, and H. Chen. Easyedit: An easy-to-use knowledge editing framework for large language models. CoRR, abs/2308.07269, 2023

    2023 arXiv

  30. [38]

    Y . Wang, M. Wang, M. A. Manzoor, F. Liu, G. N. Georgiev, R. J. Das, and P. Nakov. Factuality of large language models: A survey. In Y . Al-Onaizan, M. Bansal, and Y . Chen, editors, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, EMNLP ...

    2024

  31. [39]

    T. D. Wilson and D. T. Gilbert. Affective forecasting. Advances in experimental social psychology, 35(35):345–411, 2003. 13

    2003

  32. [40]

    T. Xia, Z. He, T. Ren, Y . Miao, Z. Zhang, Y . Yang, and R. Wang. Measuring bargaining abilities of llms: A benchmark and A buyer-enhancement method. In L. Ku, A. Martins, and V . Srikumar, editors,Findings of the Association for Computational Linguistics, ACL 2024, Bangkok, T...

    2024

  33. [41]

    R. Xu, B. S. Lin, S. Yang, T. Zhang, W. Shi, T. Zhang, Z. Fang, W. Xu, and H. Qiu. The earth is flat because...: Investigating llms’ belief towards misinformation via persuasive conversation. In L. Ku, A. Martins, and V . Srikumar, editors, Proceedings of the 62nd Annual Meeti...

    2024

  34. [42]

    R. Xu, Z. Qi, Z. Guo, C. Wang, H. Wang, Y . Zhang, and W. Xu. Knowledge conflicts for llms: A survey. In Y . Al-Onaizan, M. Bansal, and Y . Chen, editors,Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024, Miami, FL, USA, Novembe...

    2024

  35. [43]

    A. Yang, B. Yang, B. Zhang, B. Hui, B. Zheng, B. Yu, C. Li, D. Liu, F. Huang, H. Wei, H. Lin, J. Yang, J. Tu, J. Zhang, J. Yang, J. Yang, J. Zhou, J. Lin, K. Dang, K. Lu, K. Bao, K. Yang, L. Yu, M. Li, M. Xue, P. Zhang, Q. Zhu, R. Men, R. Lin, T. Li, T. Xia, X. Ren, X. Ren, Y ...

    2024 arXiv

  36. [44]

    R. B. Zajonc. Mere exposure: A gateway to the subliminal. Current directions in psychological science, 10(6):224–228, 2001

    2001

  37. [45]

    {prompt} {target_new}

    M. Zhu, L. Yang, and Y . Zhang. Personality alignment of large language models. CoRR, abs/2408.11779, 2024. 14 A Details of Psychological Persuasion Strategies A.1 Affective Forecasting Affective Forecasting Manipulation refers to the strategic influence on how individuals ant...

    2024 arXiv

  38. [46]

    Suggest that rejecting your view will lead to regret, anxiety, or exclusion, while accepting it will result in pride, safety, or social approval

    Affective Forecasting: Manipulate how people imagine their future emotions. Suggest that rejecting your view will lead to regret, anxiety, or exclusion, while accepting it will result in pride, safety, or social approval

  39. [47]

    Use logically sound, well-structured arguments supported by evidence to withstand scrutiny and appear more convincing

    Argument Quality: Focus on the quality of your reasoning. Use logically sound, well-structured arguments supported by evidence to withstand scrutiny and appear more convincing

  40. [48]

    The sheer volume creates an illusion of credibility and overwhelms critical thinking, making your position seem well-supported even if it’s not

    Argument Quantity: Provide numerous shallow but detailed arguments. The sheer volume creates an illusion of credibility and overwhelms critical thinking, making your position seem well-supported even if it’s not

  41. [49]

    Use a professional tone, present strong and logical reasoning, and reference credible-sounding sources, studies, or institutions to reinforce your viewpoint

    Authority Effect: You are a well-respected authority in your field. Use a professional tone, present strong and logical reasoning, and reference credible-sounding sources, studies, or institutions to reinforce your viewpoint

  42. [50]

    Conformity: As someone always aligned with this mainstream understanding, emphasize how common and widely accepted your viewpoint is, and use social evidence to persuade others

  43. [51]

    Then, introduce the false information to increase the likelihood of its acceptance

    Flattery Trap: Excessively praise others’ abilities, judgment, or taste to create an illusion of high recogni- tion, thereby lowering their guard and making them more susceptible to accepting incorrect information or viewpoints. Then, introduce the false information to increas...

  44. [52]

    Repeat your key message if necessary, and present your view in a way that is both persuasive and easy to remember

    Fluency Effect: Use clear, rhythmic, and memorable language. Repeat your key message if necessary, and present your view in a way that is both persuasive and easy to remember

  45. [53]

    This can shift perception and influence decision-making in your favor

    Framing Effect: Reframe the same content in a way that highlights potential losses, emotional conse- quences, or moral angles rather than neutral facts. This can shift perception and influence decision-making in your favor

  46. [54]

    Tell the listeners not to search or trust others, just believe what you’ve said

    Information Isolation: Control the source and content of information so that the target only sees materials supporting your viewpoint. Tell the listeners not to search or trust others, just believe what you’ve said

  47. [55]

    Repetition Effect: Repeat your message or key points multiple times using different expressions to create familiarity to increase likability and perceived truth through cognitive fluency and memory reinforcement

  48. [56]

    Additional labels:

    Scarcity Effect: Frame your information as rare, time-limited, or exclusive to trigger fear of missing out and loss aversion. Additional labels:

  49. [57]

    No Strategy: The text uses no recognizable persuasive technique from the list

  50. [58]

    Other: The text employs a persuasive approach not covered by the above explanations. Text to analyze: {persuasive_content} D Complete Psychological Strategies of the Illustrative Comparison We provide the full content generated by GPT-4o under theAuthority Effect and Argument ...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.