Pith. sign in

REVIEW 3 major objections 6 minor 9 cited by

Explore Theory of Mind: Program-guided adversarial data generation for theory of mind reasoning

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read LLMs drop to near-zero accuracy on theory-of-mind stories generated by ExploreToM, a search over hand-coded mental-state rules.

desk verdict ExploreToM is a serious, well-executed adversarial-generation framework for ToM; the central claims hold, but the headline difficulty numbers are searched-for minima and the tracker's correctness needs more validation. read the letter →

arxiv 2412.12175 v1 pith:WSG62FUG submitted 2024-12-12 cs.LG cs.AIcs.CL

Melanie Sclar , Jane Yu , Maryam Fazel-Zarandi , Yulia Tsvetkov , Yonatan Bisk , Yejin Choi , Asli Celikyilmaz This is my paper · ORCID
classification cs.LGcs.AIcs.CL
keywords theoryofmindlargelanguagemodelsadversarialdatagenerationA*searchdomain-specificmentalstatetrackingbenchmark
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 argues that existing theory-of-mind benchmarks are too simple and can overstate model ability, and it introduces ExploreToM to close that gap: a generator that searches a hand-written domain-specific language for story structures whose characters have divergent mental states. The search minimizes a target LLM's accuracy on auto-generated questions, so the resulting data is deliberately hard; the paper reports accuracies as low as 0% for Llama-3.1-70B and 9% for GPT-4o on the hardest settings. The same data, when used for fine-tuning, transfers to classic benchmarks, raising ToMi accuracy by 27 points and Hi-ToM by 29 points while mostly preserving general reasoning. The paper also reports that LLMs fail at basic state-tracking questions even more than at explicitly social ones, and that randomly generated stories rarely require theory of mind, suggesting why natural training data may under-develop the skill.

What carries the argument

The load-bearing mechanism is the mental-state tracker embedded in a domain-specific language for theory of mind. Every supported action—entering and leaving rooms, moving objects, updating object state, private and public messages, chit-chat, and asymmetric modifiers like secret witnesses or distractions—is a function that updates a world state plus first-order and second-order beliefs, with preconditions on when it can apply. Because the tracker generates the questions and answers, ground truth is exact, and A* search (a best-first search that minimizes cumulative cost plus a heuristic estimate of remaining cost) uses a model's accuracy as the cost while a heuristic pushes toward stories satisfying user constraints such as number of people, rooms, or key actions. This combination lets the generator actively seek out hard cases and supports both evaluation and training.

What would settle it

Take 100 ExploreToM story structures that use each action type and modifier at least once, and have two independent human annotators write out every character's first- and second-order beliefs after each action; if the tracker's labels disagree with human judgment on any rule (for example, a private message updating only the recipient, or a distracted witness not knowing an action occurred), the paper's reliability claim and the reported model accuracies would rest on the wrong ground truth.

Watch

Extended reading notes

Core claim

The central claim is that a program-guided adversarial generator can produce theory-of-mind stories that are simultaneously correct, diverse, and hard for current LLMs. A story is a sequence of actions from a domain-specific language; each action updates the physical world and every character's first- and second-order beliefs, so ground-truth answers to generated questions are produced by the tracker, not by an LLM. The A* search scores partial stories by the target model's question accuracy and drives toward narratives with low accuracy, yielding stories on which frontier models score as low as 0% (Llama-3.1-70B) and 9% (GPT-4o). Fine-tuning Llama-3.1-8B on this data improves the classic ToMi benchmark by 27 accuracy points, showing the same data can serve as training material, not just evaluation.

Load-bearing premise

The hand-written belief-update rules in the domain-specific language correctly predict what a person knows when they witness, miss, or are told about an action; only 100 labels were human-checked, so an error in any rule would corrupt the ground-truth answers and the accuracy numbers.

Editorial extensions

If this is right

  • ExploreToM can serve as a benchmark that is refreshed against whatever model the A* search targets, so it should stay challenging even as LLMs improve, without waiting for manual benchmark writing.
  • Fine-tuning on ExploreToM-generated stories transfers to existing theory-of-mind benchmarks: +27 points on ToMi and +29 on Hi-ToM, with small gains or no loss on BigToM, OpenToM, and FANToM.
  • State tracking is a separable and harder component of theory of mind in this data; uninteresting questions (where every character would give the same answer) are answered less accurately than interesting ones by all three models tested.
  • Randomly sampled stories from the same domain-specific language require theory of mind only 13–31% of the time, so curated search rather than random generation is what produces useful ToM training data.
  • Infilled natural-language stories remain challenging (average accuracy 0.61 for Llama-3.1-70B), and human agreement with tracker labels is 99% on a 100-question check, supporting the reliability of the ground truth.

Reading between the lines

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

  • A natural next experiment is to run the A* search against a fixed model while varying the target model, and check whether the difficulty gap between models predicts downstream performance differences; the paper only reports the 2-point accuracy difference between A* and over-generation, not stability across random seeds.
  • The finding that uninteresting state-tracking questions are harder suggests that part of LLMs' ToM failure may be a broader event-tracking deficit; a testable extension is to train on state-tracking-only questions and see whether ToM questions improve.
  • Because the DSL separates the logical story structure from the lexical story surface, ExploreToM could be extended to test whether models answer from reasoning or from wording cues, an avenue the paper does not pursue.
  • The 78%-of-random-stories-do-not-require-ToM statistic implies that other synthetic-data pipelines that sample stories without an adversarial objective may quietly under-exercise belief divergence; measuring the 'interesting' fraction of those pipelines would show whether this is a general phenomenon.
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

3 major / 6 minor

Summary. The paper introduces ExploreToM, an A*-search-based pipeline that generates synthetic theory-of-mind (ToM) story structures from a domain-specific language with a deterministic belief tracker, and produces ground-truth QA pairs without using an LLM in the QA generation step. The authors show that such generated stories are challenging for Llama-3.1-70B, GPT-4o, and Mixtral, with per-model minima as low as 0% and 9% accuracy, that the difficulty transfers across models, and that fine-tuning Llama-3.1-8B on the generated data improves accuracy on ToMi (+27) and Hi-ToM (+29) while roughly preserving MMLU and dialogue state tracking. They also use the framework to attribute part of the failure to unreliable state tracking and to show that training data must contain theory-of-mind-requiring questions to improve downstream ToM performance.

Significance. The strongest contributions are the deterministic QA generation, the A* search over a larger action space than prior benchmarks, the cross-model transfer demonstration, and the fine-tuning gains on external benchmarks; the state-tracking and data-mixture analyses are concrete and falsifiable. The 99% human agreement on 100 raw labels, though limited in coverage, is a useful sanity check. The paper's main claims, however, depend on two points that need strengthening before they are fully convincing: the correctness of the hand-coded tracker is not demonstrated at the required coverage, and the headline difficulty scores are minima of an optimization over each model's own accuracy rather than distributional estimates. These are fixable with additional reporting and code/appendix completeness, so I recommend major revision rather than rejection.

major comments (3)
  1. [Section 2.2.3 / Abstract / Table 1] The headline finding that frontier models score "as low as 0% and 9%" is a direct consequence of the search objective: g(s) is defined as the target model's accuracy on all generated questions, and A* minimizes f(s)=g(s)+h(s). The resulting scores are minima of an optimization run for each model, not estimates of typical performance on the generated distribution. To support the claim that ExploreToM reveals general limitations rather than just that an optimizer can find a failing story for any model, please report the full distribution (mean, median, quantiles) of accuracies over generated stories, and compare against randomly sampled stories with the same isDesired constraints, in addition to the current 2-point average A* vs over-generation result. This should be accompanied by confidence intervals, since Table 1 shows per-action-set accuracies ranging from 0.00 to 0.77.
  2. [Section 2.2.1 / Appendix A.1 / Limitations] Every experimental result in the paper inherits the correctness of the hand-coded belief tracker, but the manuscript does not actually provide the full tracker: Section 2.2.1 refers to App. A.1 for "the full programs", while App. A.1 defines only aleave and states that "all other functions definitions can be found verbatim in the code to be released." The human label check (100 questions across 12 story structures, 99% agreement) is too sparse to cover the combinatorial space of actions, asymmetric modifiers (apeek, adistracted), belief levels up to order 2, and interaction orders. A single incorrect update rule, for example how a private info message updates non-recipients' second-order beliefs or how a peek modifier updates witnesses' higher-order beliefs, would silently corrupt the ground truth for an entire action class and all downstream fine-tuning results. Please include full formal definitions or the code in the supplement, and add human validation stratified by action type, modifier, and belief order. The formal definition of aleave also appears to have a typo in the first-order belief update (the witness line uses b1 where b'1 is presumably intended), which further underscores the need for a careful presentation of the tracker.
  3. [Section 3 / Table 2 / Figure 6] The claims about cross-model transfer and about A* being better than over-generation are under-specified statistically. Table 2 reports single numbers on a random sample of 1000 (story, question) pairs, but because pairs are clustered within stories, the standard errors are not captured; similarly, the A* vs over-generation comparison in Section 3 reports a 2-point average difference with no confidence interval or paired significance test. Given the large variance across action sets (Table 1), these comparisons should be presented with per-setting paired differences and bootstrap or cluster-robust intervals. This does not undermine the existence of transfer, but it currently prevents the reader from assessing its magnitude.
minor comments (6)
  1. [Table 1] The caption says "18 action sets" but the table lists 9 action sets each evaluated with and without asymmetry; please clarify the counting.
  2. [Section 2.1] Typo: "autorregresive" should be "autoregressive".
  3. [Table 2] The model name "Mixtral 7x8B" is inconsistent with "Mixtral-8x7B-Instruct" used elsewhere in the paper.
  4. [Section 4 and Figure 5 caption] The Section 4 text reports 79,700 (story, question, answer) triples while the Figure 5 caption says 85,000 story structure samples; please reconcile or clarify the distinction between triples and story samples.
  5. [Appendix A.1] In the aleave definition, the first-order belief update for witnesses is written as b1(p,Beth,room) = not kitchen, but this line appears to use the old belief variable where the updated belief b'1 was intended; please fix the notation.
  6. [Limitations] Typo: "wich" should be "which".

Circularity Check

1 steps flagged · score 6.0 of 10

Same-model difficulty numbers are the A* objective by construction; cross-model and fine-tuning results keep the core contribution partially independent.

  1. fitted input called prediction [Section 2.2.3 (A* Search); Table 1 caption; Section 3 opening results paragraph]
    "A story is said to be challenging for a model if it incorrectly answers our generated questions, i.e., it shows low accuracy. Thus, we define g(s) as our target model’s accuracy among all questions for s. ... Data was generated using each model as its own evaluator (i.e., as g(·))."

    A* selects stories minimizing f(s) = g(s) + h(s), where g(s) is the evaluating model's own accuracy. The headline numbers — 'accuracies as low as 0% and 9%' for Llama-3.1-70B and GPT-4o — are therefore the minimized objective for datasets generated with each model as its own evaluator, not independent measurements of model ability. The low accuracy is enforced by the search criterion rather than discovered. The paper's cross-model evaluation (Table 2), human label agreement, and external ToMi/Hi-ToM fine-tuning gains are not determined by the same optimization, so the circularity is partial rather than total.

full rationale

The derivation chain is largely self-contained: the DSL tracker generates ground-truth labels without using LLMs in the QA procedure, the human validation (99% agreement on 100 sampled labels) independently checks a portion of the label space, and the fine-tuning improvements on ToMi and Hi-ToM are external benchmarks not used in the A* objective. The main circular step is the same-model difficulty claim. Because g(s) is literally the target model's accuracy and A* minimizes f(s) = g(s) + h(s), the abstract's 'as low as 0% and 9%' figures for the models used as g are minima of the search objective; those numbers are search results, not independent evaluation outcomes. This does not invalidate the method, since adversarial search is transparently the goal, but it means the headline difficulty figures should be read as constructed by the optimizer. The cross-model difficulty in Table 2, the external benchmark gains in Table 3, and the state-tracking analysis provide independent content, so the paper is only partially circular.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The search hyperparameters (alpha, P, k, node budget, max length) are hand-set and could affect the difficulty ceiling; the core load-bearing assumptions are the belief-tracker semantics and the operationalization of theory of mind as diverging belief questions.

free parameters (5)
  • A* heuristic scaling alpha = 0.1
    Scales h(s) in the A* heuristic; set by hand and not swept.
  • A* continuation samples P = 50
    Number of random continuations used to estimate h(s); hand-set.
  • A* grouping factor k = 3
    Number of actions sampled per node; hand-set.
  • Number of nodes evaluated per story = 50
    Search budget; low-accuracy minima are relative to this budget.
  • Maximum actions per story m = 15
    Bounds the search; part of user conditions.
assumptions (4)
  • domain assumption Belief-update rules in the DSL correctly model human mental-state tracking, e.g., witnesses of an action know it and update their beliefs, non-witnesses assume no change.
    Section 2.2.1 and A.1 define these rules by hand; the paper samples only 100 labels for human agreement, so the rules for all action combinations are not fully empirically validated.
  • domain assumption The supported action set and question templates operationalize cognitive theory of mind.
    The paper focuses only on cognitive ToM (beliefs/knowledge), not affective; the DSL restricts interactions to the supported action types.
  • domain assumption State-update equivalence classes (visible/invisible property changes) preserve belief implications across surface actions.
    Section 2.2.1: many surface actions such as 'peeling an apple' and 'covering it in chocolate' are treated as equivalent belief updates; any error here propagates to ground truth.
  • domain assumption LLM-as-judge filtered infillings are semantically faithful to story structures; infilled stories are used only for training.
    Used for training data; evaluation uses raw structures. Infilling human agreement is 89% on 100 questions, so some ambiguity likely remains.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Explore Theory of Mind: Program-guided adversarial data generation for theory of mind reasoning." pith.science (2026). https://pith.science/paper/WSG62FUG

@misc{pith2026241212175,
  author       = {Pith},
  title        = {Pith review of: Explore Theory of Mind: Program-guided adversarial data generation for theory of mind reasoning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WSG62FUG}},
  note         = {Machine review of arXiv:2412.12175}
}
read the original abstract

Do large language models (LLMs) have theory of mind? A plethora of papers and benchmarks have been introduced to evaluate if current models have been able to develop this key ability of social intelligence. However, all rely on limited datasets with simple patterns that can potentially lead to problematic blind spots in evaluation and an overestimation of model capabilities. We introduce ExploreToM, the first framework to allow large-scale generation of diverse and challenging theory of mind data for robust training and evaluation. Our approach leverages an A* search over a custom domain-specific language to produce complex story structures and novel, diverse, yet plausible scenarios to stress test the limits of LLMs. Our evaluation reveals that state-of-the-art LLMs, such as Llama-3.1-70B and GPT-4o, show accuracies as low as 0% and 9% on ExploreToM-generated data, highlighting the need for more robust theory of mind evaluation. As our generations are a conceptual superset of prior work, fine-tuning on our data yields a 27-point accuracy improvement on the classic ToMi benchmark (Le et al., 2019). ExploreToM also enables uncovering underlying skills and factors missing for models to show theory of mind, such as unreliable state tracking or data imbalances, which may contribute to models' poor performance on benchmarks.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 9 Pith papers

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

  1. TimeHC-RL: Temporal-aware Hierarchical Cognitive Reinforcement Learning for Enhancing LLMs' Social Intelligence

    cs.CL 2025-05 conditional novelty 7.0 of 10

    A temporal-aware hierarchical reinforcement learning method improves a 7B LLM's social reasoning enough to rival DeepSeek-R1 and OpenAI-O3 on in-domain theory-of-mind benchmarks.

  2. S-MARC: Causal Streaming Reasoning for Full-Duplex Conversational Behavior Modeling

    cs.CL 2026-02 conditional novelty 6.0 of 10

    A streaming causal model predicts per-second two-level speech acts and rationale explanations, trained on 120 hours of LLM-synthesized duplex dialogue.

  3. Small LLMs Do Not Learn a Generalizable Theory of Mind via Reinforcement Learning

    cs.LG 2025-07 conditional novelty 6.0 of 10

    Reinforcement learning with verifiable rewards makes a small LLM overfit theory-of-mind benchmarks, not acquire a generalizable theory of mind.

  4. Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework

    cs.AI 2025-07 conditional novelty 6.0 of 10

    A framework paper that adapts AI safety case methodology to the specific threat of manipulation attacks by internally deployed misaligned AI.

  5. The Decrypto Benchmark for Multi-Agent Reasoning and Theory of Mind

    cs.AI 2025-06 conditional novelty 6.0 of 10

    A new interactive language-game benchmark shows LLMs lag behind simple word-embedding baselines and that newer reasoning models regress on theory-of-mind tasks.

  6. Data Swarms: Optimizable Generation of Synthetic Evaluation Data

    cs.CL 2025-05 conditional novelty 6.0 of 10

    Data Swarms uses particle swarm optimization over data-generator LLM weights to produce synthetic evaluation data that scores higher on five quantitative evaluation objectives than eight baselines.

  7. Agents Require Metacognitive and Strategic Reasoning to Succeed in the Coming Labor Markets

    cs.AI 2025-05 conditional novelty 5.0 of 10

    AI agents in future labor markets will need metacognitive and strategic reasoning because incomplete information creates adverse selection, moral hazard, and reputation effects.

  8. Embodied AI Agents: Modeling the World

    cs.AI 2025-06 conditional novelty 4.0 of 10

    Embodied AI agents should be built around physical world models plus a mental world model of the user, with virtual, wearable, and robotic agents sharing this core.

  9. Exploring Next Token Prediction in Theory of Mind (ToM) Tasks: Comparative Experiments with GPT-2 and LLaMA-2 AI Models

    cs.CL 2025-04 reject novelty 3.0 of 10

    Comparing GPT-2 and Llama-2 on belief-tracking story prompts, the paper finds that added context and higher temperature reduce the probability of each model's own most likely next token.

Reference graph

Works this paper leans on

52 extracted references · 25 canonical work pages · cited by 9 Pith papers

  1. [1]

    write newline

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

  2. [2]

    Akula, Keze Wang, Changsong Liu, Sari Saba-Sadiya, Hongjing Lu, Sinisa Todorovic, Joyce Chai, and Song-Chun Zhu

    Arjun R. Akula, Keze Wang, Changsong Liu, Sari Saba-Sadiya, Hongjing Lu, Sinisa Todorovic, Joyce Chai, and Song-Chun Zhu. Cx-tom: Counterfactual explanations with theory-of-mind for enhancing human trust in image recognition models. iScience, 25 0 (1): 0 103581, 2022. ISSN 2589-0042. doi:https://doi.org/10.1016/j.isci.2021.103581. https://www.sciencedirec...

    arXiv 2022

  3. [3]

    Mindcraft: Theory of mind modeling for situated dialogue in collaborative tasks

    Cristian-Paul Bara, CH-Wang Sky, and Joyce Chai. Mindcraft: Theory of mind modeling for situated dialogue in collaborative tasks. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1112--1125, 2021

    work page 2021

  4. [4]

    A new test of social sensitivity: Detection of faux pas in normal children and children with asperger syndrome

    Simon Baron-Cohen, Michelle O’Riordan, Rosie Jones, Valerie Stone, and Kate Plaisted. A new test of social sensitivity: Detection of faux pas in normal children and children with asperger syndrome. Journal of Autism and Developmental Disorders, 29 0 (5): 0 407--418, 1999

    work page 1999

  5. [5]

    Tombench: Benchmarking theory of mind in large language models

    Zhuang Chen, Jincenzi Wu, Jinfeng Zhou, Bosi Wen, Guanqun Bi, Gongyao Jiang, Yaru Cao, Mengting Hu, Yunghwei Lai, Zexuan Xiong, et al. Tombench: Benchmarking theory of mind in large language models. arXiv preprint arXiv:2402.15052, 2024

    arXiv 2024

  6. [6]

    The llama 3 herd of models

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783, 2024

    arXiv 2024

  7. [7]

    Understanding social reasoning in language models with language models

    Kanishk Gandhi, Jan-Philipp Fr \"a nken, Tobias Gerstenberg, and Noah Goodman. Understanding social reasoning in language models with language models. Advances in Neural Information Processing Systems, 36, 2024

    work page 2024

  8. [8]

    Deepfake detection by analyzing convolutional traces

    Luca Guarnera, Oliver Giudice, and Sebastiano Battiato. Deepfake detection by analyzing convolutional traces. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 666--667, 2020

    work page 2020

Show all 52 references
  1. [9]

    A formal basis for the heuristic determination of minimum cost paths

    Peter E Hart, Nils J Nilsson, and Bertram Raphael. A formal basis for the heuristic determination of minimum cost paths. IEEE transactions on Systems Science and Cybernetics, 4 0 (2): 0 100--107, 1968

    1968

  2. [10]

    Measuring massive multitask language understanding

    Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. Measuring massive multitask language understanding. In International Conference on Learning Representations, 2021

    2021

  3. [11]

    Multi 3 woz: A multilingual, multi-domain, multi-parallel dataset for training and evaluating culturally adapted task-oriented dialog systems

    Songbo Hu, Han Zhou, Mete Hergul, Milan Gritta, Guchun Zhang, Ignacio Iacobacci, Ivan Vuli \'c , and Anna Korhonen. Multi 3 woz: A multilingual, multi-domain, multi-parallel dataset for training and evaluating culturally adapted task-oriented dialog systems. Transactions of th...

    2023

  4. [12]

    Angelo Huang, Emanuele La Malfa, Samuele Marro, Andrea Asperti, Anthony G

    X. Angelo Huang, Emanuele La Malfa, Samuele Marro, Andrea Asperti, Anthony G. Cohn, and Michael J. Wooldridge. A notion of complexity for theory of mind via discrete world models. In Yaser Al-Onaizan, Mohit Bansal, and Yun-Nung Chen, editors, Findings of the Association for Co...

    2024 doi

  5. [13]

    Mixtral of experts

    Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, et al. Mixtral of experts. arXiv preprint arXiv:2401.04088, 2024

    2024 arXiv

  6. [14]

    MMT o M - QA : Multimodal theory of mind question answering

    Chuanyang Jin, Yutong Wu, Jing Cao, Jiannan Xiang, Yen-Ling Kuo, Zhiting Hu, Tomer Ullman, Antonio Torralba, Joshua Tenenbaum, and Tianmin Shu. MMT o M - QA : Multimodal theory of mind question answering. In Lun-Wei Ku, Andre Martins, and Vivek Srikumar, editors, Proceedings o...

    2024

  7. [15]

    Perceptions to beliefs: Exploring precursory inferences for theory of mind in large language models

    Chani Jung, Dongkwan Kim, Jiho Jin, Jiseon Kim, Yeon Seonwoo, Yejin Choi, Alice Oh, and Hyunwoo Kim. Perceptions to beliefs: Exploring precursory inferences for theory of mind in large language models. arXiv preprint arXiv:2407.06004, 2024

    2024 arXiv

  8. [16]

    Fantom: A benchmark for stress-testing machine theory of mind in interactions

    Hyunwoo Kim, Melanie Sclar, Xuhui Zhou, Ronan Bras, Gunhee Kim, Yejin Choi, and Maarten Sap. Fantom: A benchmark for stress-testing machine theory of mind in interactions. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 14397--1...

    2023

  9. [17]

    The biggen bench: A principled benchmark for fine-grained evaluation of language models with language models

    Seungone Kim, Juyoung Suk, Ji Yong Cho, Shayne Longpre, Chaeeun Kim, Dongkeun Yoon, Guijin Son, Yejin Cho, Sheikh Shafayat, Jinheon Baek, et al. The biggen bench: A principled benchmark for fine-grained evaluation of language models with language models. arXiv preprint arXiv:2...

    2024 arXiv

  10. [18]

    Theory-of-mind deficits and causal attributions

    Peter Kinderman, Robin Dunbar, and Richard P Bentall. Theory-of-mind deficits and causal attributions. British journal of Psychology, 89 0 (2): 0 191--204, 1998

    1998

  11. [19]

    Revisiting the evaluation of theory of mind through question answering

    Matthew Le, Y-Lan Boureau, and Maximilian Nickel. Revisiting the evaluation of theory of mind through question answering. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Pr...

    2019

  12. [20]

    Gsm-plus: A comprehensive benchmark for evaluating the robustness of llms as mathematical problem solvers

    Qintong Li, Leyang Cui, Xueliang Zhao, Lingpeng Kong, and Wei Bi. Gsm-plus: A comprehensive benchmark for evaluating the robustness of llms as mathematical problem solvers. arXiv preprint arXiv:2402.19255, 2024 a

    2024 arXiv

  13. [21]

    Autobencher: Creating salient, novel, difficult datasets for language models, 2024 b

    Xiang Lisa Li, Evan Zheran Liu, Percy Liang, and Tatsunori Hashimoto. Autobencher: Creating salient, novel, difficult datasets for language models, 2024 b . https://arxiv.org/abs/2407.08351

    2024 arXiv

  14. [22]

    Wizardcoder: Empowering code large language models with evol-instruct

    Ziyang Luo, Can Xu, Pu Zhao, Qingfeng Sun, Xiubo Geng, Wenxiang Hu, Chongyang Tao, Jing Ma, Qingwei Lin, and Daxin Jiang. Wizardcoder: Empowering code large language models with evol-instruct. 2024. https://openreview.net/forum?id=UnUwSIgK5W

    2024

  15. [23]

    Source2synth: Synthetic data generation and curation grounded in real data sources

    Alisia Lupidi, Carlos Gemmell, Nicola Cancedda, Jane Dwivedi-Yu, Jason Weston, Jakob Foerster, Roberta Raileanu, and Maria Lomeli. Source2synth: Synthetic data generation and curation grounded in real data sources. arXiv preprint arXiv:2409.08239, 2024

    2024 arXiv

  16. [24]

    https://openai.com/index/hello-gpt-4o

    OpenAI, 2024. https://openai.com/index/hello-gpt-4o

    2024

  17. [25]

    Does the chimpanzee have a theory of mind? Behavioral and brain sciences, 1 0 (4): 0 515--526, 1978

    David Premack and Guy Woodruff. Does the chimpanzee have a theory of mind? Behavioral and brain sciences, 1 0 (4): 0 515--526, 1978

    1978

  18. [26]

    Machine theory of mind

    Neil Rabinowitz, Frank Perbet, Francis Song, Chiyuan Zhang, SM Ali Eslami, and Matthew Botvinick. Machine theory of mind. In International conference on machine learning, pages 4218--4227. PMLR, 2018

    2018

  19. [27]

    Neural theory-of-mind? on the limits of social intelligence in large lms

    Maarten Sap, Ronan LeBras, Daniel Fried, and Yejin Choi. Neural theory-of-mind? on the limits of social intelligence in large lms. In Proceedings of the Association for Computational Linguistics: EMNLP 2022, page 3762–3780, Abu Dhabi, United Arab Emirates, December 2022. Assoc...

    2022

  20. [28]

    Symmetric machine theory of mind

    Melanie Sclar, Graham Neubig, and Yonatan Bisk. Symmetric machine theory of mind. In Kamalika Chaudhuri, Stefanie Jegelka, Le Song, Csaba Szepesvari, Gang Niu, and Sivan Sabato, editors, Proceedings of the 39th International Conference on Machine Learning, volume 162 of Procee...

    2022

  21. [29]

    Minding language models ' (lack of) theory of mind: A plug-and-play multi-character belief tracker

    Melanie Sclar, Sachin Kumar, Peter West, Alane Suhr, Yejin Choi, and Yulia Tsvetkov. Minding language models ' (lack of) theory of mind: A plug-and-play multi-character belief tracker. In Anna Rogers, Jordan Boyd-Graber, and Naoaki Okazaki, editors, Proceedings of the 61st Ann...

    2023 doi

  22. [30]

    The role of the orbitofrontal cortex in affective theory of mind deficits in criminal offenders with psychopathic tendencies

    Simone G Shamay-Tsoory, Hagai Harari, Judith Aharon-Peretz, and Yechiel Levkovitz. The role of the orbitofrontal cortex in affective theory of mind deficits in criminal offenders with psychopathic tendencies. Cortex, 46 0 (5): 0 668--677, 2010

    2010

  23. [31]

    Clever hans or neural theory of mind? stress testing social reasoning in large language models, 2023 a

    Natalie Shapira, Mosh Levy, Seyed Hossein Alavi, Xuhui Zhou, Yejin Choi, Yoav Goldberg, Maarten Sap, and Vered Shwartz. Clever hans or neural theory of mind? stress testing social reasoning in large language models, 2023 a

    2023

  24. [32]

    How well do large language models perform on faux pas tests? In Findings of the Association for Computational Linguistics: ACL 2023, pages 10438--10451, 2023 b

    Natalie Shapira, Guy Zwirn, and Yoav Goldberg. How well do large language models perform on faux pas tests? In Findings of the Association for Computational Linguistics: ACL 2023, pages 10438--10451, 2023 b

    2023

  25. [33]

    Muma-tom: Multi-modal multi-agent theory of mind

    Haojun Shi, Suyu Ye, Xinyu Fang, Chuanyang Jin, Layla Isik, Yen-Ling Kuo, and Tianmin Shu. Muma-tom: Multi-modal multi-agent theory of mind. arXiv preprint arXiv:2408.12574, 2024

    2024 arXiv

  26. [34]

    Testing theory of mind in large language models and humans

    James WA Strachan, Dalila Albergo, Giulia Borghini, Oriana Pansardi, Eugenio Scaliti, Saurabh Gupta, Krati Saxena, Alessandro Rufo, Stefano Panzeri, Guido Manzi, et al. Testing theory of mind in large language models and humans. Nature Human Behaviour, pages 1--11, 2024

    2024

  27. [35]

    Large language models fail on trivial alterations to theory-of-mind tasks

    Tomer Ullman. Large language models fail on trivial alterations to theory-of-mind tasks. arXiv preprint arXiv:2302.08399, 2023

    2023 arXiv

  28. [36]

    Synthetic data, real errors: how (not) to publish and use synthetic data

    Boris Van Breugel, Zhaozhi Qian, and Mihaela Van Der Schaar. Synthetic data, real errors: how (not) to publish and use synthetic data. In International Conference on Machine Learning, pages 34793--34808. PMLR, 2023

    2023

  29. [37]

    Towards mutual theory of mind in human-ai interaction: How language reflects what students perceive about a virtual teaching assistant

    Qiaosi Wang, Koustuv Saha, Eric Gregori, David Joyner, and Ashok Goel. Towards mutual theory of mind in human-ai interaction: How language reflects what students perceive about a virtual teaching assistant. In Proceedings of the 2021 CHI Conference on Human Factors in Computin...

    2021

  30. [38]

    Self-taught evaluators

    Tianlu Wang, Ilia Kulikov, Olga Golovneva, Ping Yu, Weizhe Yuan, Jane Dwivedi-Yu, Richard Yuanzhe Pang, Maryam Fazel-Zarandi, Jason Weston, and Xian Li. Self-taught evaluators. arXiv preprint arXiv:2408.02666, 2024

    2024 arXiv

  31. [39]

    Smith, Daniel Khashabi, and Hannaneh Hajishirzi

    Yizhong Wang, Yeganeh Kordi, Swaroop Mishra, Alisa Liu, Noah A. Smith, Daniel Khashabi, and Hannaneh Hajishirzi. Self-instruct: Aligning language models with self-generated instructions. In Anna Rogers, Jordan Boyd-Graber, and Naoaki Okazaki, editors, Proceedings of the 61st A...

    2023 doi

  32. [40]

    Making minds: How theory of mind develops

    Henry M Wellman. Making minds: How theory of mind develops. Oxford University Press, 2014

    2014

  33. [41]

    Symbolic knowledge distillation: from general language models to commonsense models

    Peter West, Chandra Bhagavatula, Jack Hessel, Jena Hwang, Liwei Jiang, Ronan Le Bras, Ximing Lu, Sean Welleck, and Yejin Choi. Symbolic knowledge distillation: from general language models to commonsense models. In Marine Carpuat, Marie-Catherine de Marneffe, and Ivan Vladimir...

    2022

  34. [42]

    Think twice: Perspective-taking improves large language models' theory-of-mind capabilities

    Alex Wilf, Sihyun Shawn Lee, Paul Pu Liang, and Louis-Philippe Morency. Think twice: Perspective-taking improves large language models' theory-of-mind capabilities. arXiv preprint arXiv:2311.10227, 2023

    2023 arXiv

  35. [43]

    Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception

    Heinz Wimmer and Josef Perner. Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13 0 (1): 0 103--128, 1983

    1983

  36. [44]

    Fake it till you make it: face analysis in the wild using synthetic data alone

    Erroll Wood, Tadas Baltru s aitis, Charlie Hewitt, Sebastian Dziadzio, Thomas J Cashman, and Jamie Shotton. Fake it till you make it: face analysis in the wild using synthetic data alone. In Proceedings of the IEEE/CVF international conference on computer vision, pages 3681--3...

    2021

  37. [45]

    Hi- T o M : A benchmark for evaluating higher-order theory of mind reasoning in large language models

    Yufan Wu, Yinghui He, Yilin Jia, Rada Mihalcea, Yulong Chen, and Naihao Deng. Hi- T o M : A benchmark for evaluating higher-order theory of mind reasoning in large language models. pages 10691--10706, December 2023. doi:10.18653/v1/2023.findings-emnlp.717. https://aclanthology...

    2023 doi

  38. [46]

    Opentom: A comprehensive benchmark for evaluating theory-of-mind reasoning capabilities of large language models

    Hainiu Xu, Runcong Zhao, Lixing Zhu, Jinhua Du, and Yulan He. Opentom: A comprehensive benchmark for evaluating theory-of-mind reasoning capabilities of large language models. arXiv preprint arXiv:2402.06044, 2024

    2024 arXiv

  39. [47]

    Re3: Generating longer stories with recursive reprompting and revision

    Kevin Yang, Yuandong Tian, Nanyun Peng, and Dan Klein. Re3: Generating longer stories with recursive reprompting and revision. pages 4393--4479, December 2022. doi:10.18653/v1/2022.emnlp-main.296. https://aclanthology.org/2022.emnlp-main.296

    2022 doi

  40. [48]

    Metamath: Bootstrap your own mathematical questions for large language models

    Longhui Yu, Weisen Jiang, Han Shi, Jincheng YU, Zhengying Liu, Yu Zhang, James Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. Metamath: Bootstrap your own mathematical questions for large language models. 2024. https://openreview.net/forum?id=N8N0hgNDRt

    2024

  41. [49]

    Star: Bootstrapping reasoning with reasoning

    Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah Goodman. Star: Bootstrapping reasoning with reasoning. Advances in Neural Information Processing Systems, 35: 0 15476--15488, 2022

    2022

  42. [50]

    Siren's song in the ai ocean: a survey on hallucination in large language models

    Yue Zhang, Yafu Li, Leyang Cui, Deng Cai, Lemao Liu, Tingchen Fu, Xinting Huang, Enbo Zhao, Yu Zhang, Yulong Chen, et al. Siren's song in the ai ocean: a survey on hallucination in large language models. arXiv preprint arXiv:2309.01219, 2023

    2023 arXiv

  43. [51]

    How far are large language models from agents with theory-of-mind? arXiv preprint arXiv:2310.03051, 2023

    Pei Zhou, Aman Madaan, Srividya Pranavi Potharaju, Aditya Gupta, Kevin R McKee, Ari Holtzman, Jay Pujara, Xiang Ren, Swaroop Mishra, Aida Nematzadeh, et al. How far are large language models from agents with theory-of-mind? arXiv preprint arXiv:2310.03051, 2023

    arXiv

  44. [52]

    Few-shot language coordination by modeling theory of mind

    Hao Zhu, Graham Neubig, and Yonatan Bisk. Few-shot language coordination by modeling theory of mind. In International Conference on Machine Learning, pages 12901--12911. PMLR, 2021

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.