Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

Meta-Thinking in LLMs via Multi-Agent Reinforcement Learning: A Survey

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper argues that making large language models truly introspective requires moving beyond single-agent self-feedback to multi-agent reinforcement learning, and it provides a systematic roadmap for that shift.

desk verdict A useful but uneven survey of meta-thinking in LLMs; the central firstness claim is not supported by its own comparison table, and the taxonomy and roadmap are the real value. read the letter →

arxiv 2504.14520 v1 pith:AHXPB7YQ submitted 2025-04-20 cs.AI cs.CL

Ahsan Bilal , Muhammad Ahmed Mohsin , Muhammad Umer , Muhammad Awais Khan Bangash , Muhammad Ali Jamshed This is my paper · ORCID
classification cs.AIcs.CL
keywords meta-thinkingmulti-agentreinforcementlearninglargelanguagemodelsself-reflectionhallucinationmitigationtheoryofmindrewarddesignsurvey
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 survey sets out to establish that meta-thinking—a model's ability to reflect on, assess, and steer its own reasoning—is the natural next step for making large language models reliable, and that multi-agent reinforcement learning (MARL) is the most promising route to install it. It argues that current fixes for hallucination, such as chain-of-thought prompting, RLHF, self-distillation, and self-feedback, all inherit the autoregressive one-token-at-a-time generation loop and lack a genuine self-checking mechanism. The paper's contribution is a roadmap: a taxonomy of single-agent, multi-agent, and reinforcement-based meta-thinking methods, a study of MARL strategies (meta-rewards, self-play and adversarial debate, meta-learning for continual adaptation), and a consolidated view of evaluation metrics and datasets. A sympathetic reader would take the paper's core thesis to be that supervisor–worker hierarchies, adversarial debates, and theory-of-mind configurations can give LLMs human-like introspection, and that pursuing this direction is necessary for high-stakes applications.

What carries the argument

The central organizing device is a five-axis comparison table (Table I) that codes prior surveys on meta-thinking, multi-agent design, RL/Meta-RL, evaluation benchmarks, and emerging directions; the paper's claim to fill a gap rests on this coding, since it says no earlier survey checks all five axes. Inside the argument, the main mechanism is a three-part MARL strategy set: a combined meta-reward $R_t=\lambda r^e_t+(1-\lambda)r^i_t$ balancing extrinsic human and task signals with intrinsic self-generated signals such as novelty or contradiction detection; collaborative self-play and adversarial training in which agents argue, critique, or attack one another's reasoning; and meta-learning loops that adapt quickly to new tasks. A named supporting architecture is the supervisor-agent hierarchy, where a high-level agent uses theory of mind to decompose tasks and revise strategies from low-level feedback.

What would settle it

Reread the Table I entries with an explicit rubric, asking of each work whether it substantively covers meta-cognition, multi-agent design, reinforcement frameworks, datasets, and emerging architectures; if any one of the 13, for instance the meta-reasoning survey [51], satisfies all five under that rubric, the paper's central positioning fails. A second, independent check would be to reproduce the ReMA result [43] on StrategyQA: if a supervisor-worker MARL system does not beat a single-agent baseline on coherence and contradiction counts, the roadmap's flagship example does not carry.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is the synthesis itself: it is, the authors say, the first survey to examine the intersection of meta-reasoning and meta-thinking in LLMs with Meta-RL and multi-agent systems. The paper organizes the field with a taxonomy into three families—single-agent methods (self-distillation, reflective prompting, chain-of-thought), multi-agent architectures (supervisor-agent hierarchies, debate and self-critique, role-play), and emerging reinforcement-based self-improvement (RLHF, adaptive self-rewarding systems)—and then argues that MARL strategies, especially intrinsic/extrinsic meta-rewards of the form $R_t=\lambda r^e_t+(1-\lambda)r^i_t$, self-play with adversarial critique, and meta-learning for fast adaptation, are what turn static LLMs into introspective, adaptive, trustworthy systems. It also assembles a set of evaluation metrics (Error Localization Accuracy, depth-wise accuracy, meta- vs. object-level accuracy, AIA/MIA, and a proposed SEQ summarization metric) and datasets (BIG-Bench, SciInstruct, DebateQA, StrategyQA, MR-Ben, FRANKLIN, Multi-LogiEval, MalAlgoQA, METAL) as a shared reference for measuring meta-thinking.

Load-bearing premise

The load-bearing premise is that Table I's binary coding of 13 earlier works—with no disclosed rubric and no independent audit—correctly shows that none of them jointly covers all five axes; if that coding is wrong, the paper's first-to-cover-everything position collapses.

Editorial extensions

If this is right

  • If the survey's synthesis is adopted, research on LLM reliability gains a shared vocabulary: single-agent self-feedback, multi-agent debate, and reinforcement-based self-improvement are treated as one design space rather than separate literatures.
  • The meta-reward formula $R_t=\lambda r^e_t+(1-\lambda)r^i_t$ gives a concrete training objective that any lab could instantiate by pairing human preference data with an LLM-as-judge intrinsic signal.
  • Supervisor-worker hierarchies and adversarial debates, if they work as described, offer a direct route to hallucination reduction: errors are caught by a second agent before the final response is emitted.
  • The metrics the survey catalogues—ELA, depth-wise accuracy, and AIA/MIA—provide a ready-made battery for testing whether a model is actually reflecting rather than repeating.
  • Following the roadmap leads to a concrete engineering target: LLMs that can flag uncertainty, revise their reasoning, and adapt to new domains without full retraining.

Reading between the lines

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

  • The paper does not test this, but its taxonomy predicts that on the AIA/MIA metrics a MARL-trained multi-agent system should improve MIA (spotting flawed reasoning) more than AIA (choosing the correct rationale), because adversarial training targets error detection specifically.
  • The five-axis comparison could be turned into a selection protocol: choose a benchmark from each axis before designing a meta-thinking system, so that coverage of the claimed gap is checked per-project rather than assumed.
  • If the meta-reward formulation is right, then reward hacking is the main risk to watch: a model could manufacture small mistakes to 'correct' them, so the survey's proposed intrinsic signals need to reward genuine coherence changes, not correction events.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The manuscript is a survey paper arguing that meta-thinking in LLMs—self-reflection, assessment, and control of reasoning—can be advanced through multi-agent reinforcement learning (MARL). It proposes a taxonomy of single-agent and multi-agent meta-thinking methods, reviews MARL strategies such as reward shaping, self-play, and meta-learning, compiles evaluation metrics and datasets, and closes with challenges and a research roadmap. Its central claim, stated in the abstract and Section I, is that this is the first systematic investigation of the intersection of Meta-RL and meta-thinking in LLMs, supported by a comparative table (Table I) intended to show that no prior survey jointly covers five axes: meta-cognition, multi-agent design, reinforcement frameworks, datasets, and emerging architectures. The paper also claims to propose a new metric, SEQ, for summarization evaluation.

Significance. If its central claims were fully substantiated, this survey would serve as a useful organizing reference for a young and rapidly growing field. The taxonomy in Section III, the reward formulation in Section IV, and the compilation of recent metrics (ELA, AIA/MIA, depth-wise accuracy) are genuinely helpful starting points, and the roadmap in Section VI identifies plausible research directions. The paper also deserves credit for attempting a structured comparison of prior surveys rather than a purely narrative review. However, the significance is conditional: the firstness claim rests almost entirely on Table I, which is not auditable as presented, and the proposed SEQ metric is an underspecified list of criteria rather than a defined metric. These are load-bearing weaknesses that must be addressed before the survey's positioning can be accepted.

major comments (4)
  1. [Section I, Contribution 3 / Table I] The firstness claim depends on Table I, but Table I is not a verifiable basis for it. The table is introduced as a comparison of survey papers, yet it includes non-survey system papers such as MetaGPT [35] and Sayed et al. [53]; the 'Meta-Thinking' column is checked for every row, including hallucination-review [10] and continual-learning surveys [47]-[49], [54], with no operational definition of the axis; and the final row 'Ahsan et al. (2025)' has no reference. A reader cannot audit any of the binary entries, and a re-coding with explicit definitions could overturn the conclusion that no prior survey covers all five axes. The authors should either supply a coding rubric with inclusion criteria, restrict the table to actual surveys, and reference their own row externally, or soften the firstness claim to what the table can demonstrably support.
  2. [Section V.A / Contribution 4] The proposed SEQ metric is not actually specified. The text describes SEQ only as measuring 'the quality of the generated summary' through five qualitative components and says it is 'derived from the METAL dataset,' but it gives no scoring formula, no aggregation method, and no validation. METAL is a multilingual meta-evaluation dataset, so the statement that a metric is 'derived from' it is unexplained. As written, SEQ is a named list of criteria, not a proposed metric; either define it precisely and provide evidence, or remove it from the contributions list.
  3. [Section IV / Contribution 2] The survey claims a 'systematic study of MARL paradigms for meta-reasoning,' but Section IV does not engage with the core machinery of MARL: there is no discussion of training paradigms such as centralized training with decentralized execution, credit assignment, communication learning, or non-stationarity, and the cited works are mostly role-based LLM agent interactions plus single-agent RL equations. The paper would either need a subsection or table that maps established MARL frameworks to the meta-thinking applications, or the title and contribution should be reframed as 'LLM-agent collaboration inspired by RL' rather than a survey of MARL.
  4. [Section I, Figure 1] Figure 1 is cited to support the claim of rapidly growing RL interest, but the figure's caption says 'Annual number of RL publications in AI conferences (2019-2024)' while the x-axis ranges from 2005 to 2020, and the y-axis label misspells 'Number.' No data source or counting methodology is given. This should be corrected or the figure should be removed, since as it stands it does not support the growth statement.
minor comments (4)
  1. [Section IV.A, Eqs. (1)-(4)] The meta-reward in Eq. (1) and the policy objective in Eq. (4) use the raw extrinsic reward r^e_t, while Eq. (2) introduces a batch-normalized version; the relationship between the normalized and raw rewards should be clarified so the notation is consistent.
  2. [Throughout] The manuscript contains many typos and running-text errors, including 'agent debase' instead of 'agent debate,' 'Continous' instead of 'Continuous,' 'accomolate' instead of 'accommodate,' and 'symbolic-MAL hybrids' for 'symbolic-MARL hybrids'; these should be corrected in a careful copyedit.
  3. [Section III.C] The statement that 'OpenAI's ChatGPT-4 is trained using RLHF' is presented as a settled fact, but the exact training procedure is not published at that level of detail; the survey should phrase this more cautiously or cite a primary source that makes the claim.
  4. [Section V.C / Figure 5] Figure 5 reports the number of published papers referencing each dataset, but no source, search method, or date of the count is given; the claim about relative popularity should either be supported by a reproducible method or described qualitatively.

Circularity Check

1 steps flagged · score 3.0 of 10

Firstness claim rests on the authors' own all-check row in Table I; no equation-level circularity elsewhere.

  1. self definitional [Section I, Contribution 3; Table I (page 5)]
    "A comparative survey of recent survey papers on LLM meta-reasoning, to show that none of the papers jointly address all five axes: 1) meta-cognition, 2) multi-agent design, 3) reinforcement frameworks, 4) existing datasets, and 5) emerging architectures. Our work fills this gap and highlights convergence points and research gaps. ... In Table I the final row reads 'Ahsan et al. (2025) ✓ ✓ ✓ ✓ ✓'."

    The uniqueness/firstness conclusion is not derived from an external benchmark; it is read off Table I, whose last row is the present paper, checked by the authors on all five self-chosen axes. The axes themselves are introduced in the same contribution as the gap claim, with no coding rubric or inclusion criteria, and the comparison set includes non-survey systems such as MetaGPT [35]. So 'no prior survey jointly covers all five axes' reduces to 'the authors did not mark any other row with all checkmarks,' and 'our work fills the gap' is the same self-assessment entered as the table's final row. This is a self-referential, self-definitional basis for the central claim, though not a mathematical equivalence.

full rationale

The paper contains no original derivation whose output is equivalent to its input. Equations (1)-(4) and the MAML inner/outer loop are standard definitions or restatements of MARL and meta-RL objectives, not claimed predictions. The self-citations [3], [6], and [122] are peripheral and not load-bearing. The taxonomy, dataset summaries, and research-challenge list are independent literature organization and remain useful regardless of the firstness claim. The only circularity-adjacent step is the firstness claim's reliance on Table I, where the authors' own all-check row is the evidence for the gap that the paper says it fills. Because the substantive survey content is externally grounded in cited primary works, this is a mild self-referential burden rather than a forced derivation.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The survey makes no empirical claims, so there are no fitted free parameters. Its central roadmap rests on a small set of domain assumptions about LLM limitations and MARL benefits, plus one paper-specific assumption about the accuracy of its own comparison table. The only invented entity is the untested SEQ metric.

assumptions (4)
  • domain assumption LLMs lack an inherent self-checking mechanism, making meta-thinking a necessary missing capability.
    Stated in Sections I and II-B as motivation; the survey does not prove that current architectures cannot acquire self-checking end-to-end.
  • domain assumption Multi-agent reinforcement learning can provide the missing meta-thinking capability for LLMs.
    Central roadmap premise. Section IV assumes agent interaction, self-play, and meta-rewards transfer to LLMs based on selected examples rather than a systematic proof.
  • ad hoc to paper Table I correctly represents prior surveys across the five claimed axes.
    The firstness claim in Contribution 3 rests entirely on this table, but no coding rubric or independent audit is provided, and MetaGPT [35], a system paper, is included as a survey.
  • domain assumption Human metacognition is an appropriate blueprint for LLM self-assessment.
    Sections II-A and VI-B propose human-like introspection and neuroscience-inspired architectures as the target; this is a design assumption, not a proven result.
invented entities (1)
  • Summary Evaluation Quality (SEQ) metric
    purpose: Proposed as a universal summarization meta-evaluation metric derived from the METAL dataset.
    Defined in Section V-A only as five bullet components: linguistic acceptability, task quality, output content quality, absence of hallucinations, and absence of problematic content. No aggregation formula, validation experiments, or comparison to established summarization metrics is given, so there is no external falsifiable handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Meta-Thinking in LLMs via Multi-Agent Reinforcement Learning: A Survey." pith.science (2026). https://pith.science/paper/AHXPB7YQ

@misc{pith2026250414520,
  author       = {Pith},
  title        = {Pith review of: Meta-Thinking in LLMs via Multi-Agent Reinforcement Learning: A Survey},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AHXPB7YQ}},
  note         = {Machine review of arXiv:2504.14520}
}
read the original abstract

This survey explores the development of meta-thinking capabilities in Large Language Models (LLMs) from a Multi-Agent Reinforcement Learning (MARL) perspective. Meta-thinking self-reflection, assessment, and control of thinking processes is an important next step in enhancing LLM reliability, flexibility, and performance, particularly for complex or high-stakes tasks. The survey begins by analyzing current LLM limitations, such as hallucinations and the lack of internal self-assessment mechanisms. It then talks about newer methods, including RL from human feedback (RLHF), self-distillation, and chain-of-thought prompting, and each of their limitations. The crux of the survey is to talk about how multi-agent architectures, namely supervisor-agent hierarchies, agent debates, and theory of mind frameworks, can emulate human-like introspective behavior and enhance LLM robustness. By exploring reward mechanisms, self-play, and continuous learning methods in MARL, this survey gives a comprehensive roadmap to building introspective, adaptive, and trustworthy LLMs. Evaluation metrics, datasets, and future research avenues, including neuroscience-inspired architectures and hybrid symbolic reasoning, are also discussed.

Figures

Figures reproduced from arXiv: 2504.14520 by the authors.

Figure 1
Figure 1. Annual number of RL publications in AI conferences (2019–2024). [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The diagram illustrates a multi-agent system where a high-level agent breaks down tasks and communicates with low-level agents to execute them. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. A COT flowchart adapted from the arithmetic examples in Wei [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Overview of RL Techniques Enabling Meta-Thinking in Language Models [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Number of published papers referencing each dataset for evaluating [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Verifying Meta-Awareness via Predictive Rewards in Reasoning Models

    cs.LG 2025-09 conditional novelty 6.0 of 10

    Rewarding reasoning models for accurately predicting their own rollout length, pass-rate, and math notions improves math benchmark accuracy and speeds up GRPO training.

Reference graph

Works this paper leans on

122 extracted references · 6 canonical work pages · cited by 1 Pith paper

  1. [75]

    Ad-autogpt: an autonomous gpt for alzheimer’s disease infodemiology,

    H. Dai, Y . Li, Z. Liu, L. Zhao, Z. Wu, S. Song, Y . Shen, D. Zhu, X. Li, S. Li, et al. , “Ad-autogpt: an autonomous gpt for alzheimer’s disease infodemiology,” arXiv preprint arXiv:2306.10095 , 2023

    arXiv 2023

  2. [35]

    Metagpt: Meta programming for multi-agent collaborative framework,

    S. Hong, X. Zheng, J. Chen, Y . Cheng, J. Wang, C. Zhang, Z. Wang, S. K. S. Yau, Z. Lin, L. Zhou, et al. , “Metagpt: Meta programming for multi-agent collaborative framework,” arXiv preprint arXiv:2308.00352, vol. 3, no. 4, p. 6, 2023

    arXiv 2023

  3. [53]

    Gizaml: A collaborative meta-learning based framework using llm for automated time-series forecasting

    E. Sayed, M. Maher, O. Sedeek, A. Eldamaty, A. Kamel, and R. El Shawi, “Gizaml: A collaborative meta-learning based framework using llm for automated time-series forecasting.” in EDBT, 2024, pp. 830–833

    2024

  4. [10]

    A review of methods for alleviating hallucination issues in large language models,

    Z. Yin, “A review of methods for alleviating hallucination issues in large language models,” Applied and Computational Engineering , vol. 76, pp. 258–266, 2024

    2024

  5. [47]

    Con- tinual learning for large language models: A survey,

    T. Wu, L. Luo, Y .-F. Li, S. Pan, T.-T. Vu, and G. Haffari, “Con- tinual learning for large language models: A survey,” arXiv preprint arXiv:2402.01364, 2024

    arXiv 2024

  6. [49]

    Towards incremental learning in large language models: A critical review,

    M. Jovanovic and P. V oss, “Towards incremental learning in large language models: A critical review,” arXiv preprint arXiv:2404.18311, 2024

    arXiv 2024

  7. [54]

    Towards lifelong learning of large language models: A survey,

    J. Zheng, S. Qiu, C. Shi, and Q. Ma, “Towards lifelong learning of large language models: A survey,” ACM Computing Surveys , vol. 57, no. 8, pp. 1–35, 2025

    2025

  8. [1]

    Is creativity without intelligence possible? a necessary condition analysis,

    M. Karwowski, J. Dul, J. Gralewski, E. Jauk, D. M. Jankowska, A. Gajda, M. H. Chruszczewski, and M. Benedek, “Is creativity without intelligence possible? a necessary condition analysis,” Intelligence, vol. 57, pp. 105–117, 2016

    2016

Show all 122 references
  1. [2]

    Thinking llms: General instruction following with thought generation,

    T. Wu, J. Lan, W. Yuan, J. Jiao, J. Weston, and S. Sukhbaatar, “Thinking llms: General instruction following with thought generation,” arXiv preprint arXiv:2410.10630, 2024

    2024 arXiv

  2. [3]

    Llms for explainable ai: A compre- hensive survey,

    A. Bilal, D. Ebert, and B. Lin, “Llms for explainable ai: A compre- hensive survey,” arXiv preprint arXiv:2504.00125 , 2025

    2025 arXiv

  3. [4]

    Chain-of-thought prompting elicits reasoning in large language models,

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. Chi, Q. V . Le, D. Zhou, et al., “Chain-of-thought prompting elicits reasoning in large language models,” Advances in neural information processing systems , vol. 35, pp. 24 824–24 837, 2022

    2022

  4. [5]

    To cot or not to cot? chain-of-thought helps mainly on math and symbolic reasoning,

    Z. Sprague, F. Yin, J. D. Rodriguez, D. Jiang, M. Wadhwa, P. Singhal, X. Zhao, X. Ye, K. Mahowald, and G. Durrett, “To cot or not to cot? chain-of-thought helps mainly on math and symbolic reasoning,” arXiv preprint arXiv:2409.12183, 2024

    2024 arXiv

  5. [6]

    Retrieval augmented generation with multi-modal llm framework for wireless environments,

    M. A. Mohsin, A. Bilal, S. Bhattacharya, and J. M. Cioffi, “Retrieval augmented generation with multi-modal llm framework for wireless environments,” arXiv preprint arXiv:2503.07670 , 2025

    2025 arXiv

  6. [7]

    A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions,

    L. Huang, W. Yu, W. Ma, W. Zhong, Z. Feng, H. Wang, Q. Chen, W. Peng, X. Feng, B. Qin, et al. , “A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions,” ACM Transactions on Information Systems , vol. 43, no. 2, pp. 1–55, 2025

    2025

  7. [8]

    Assessing llms for high stakes applications,

    S. K. Gallagher, J. Ratchford, T. Brooks, B. P. Brown, E. Heim, W. R. Nichols, S. Mcmillan, S. Rallapalli, C. J. Smith, N. VanHoudnos, et al., “Assessing llms for high stakes applications,” in Proceedings of the 46th International Conference on Software Engineering: Software E...

    2024

  8. [9]

    Siren’s song in the ai ocean: A survey on hal- lucination in large language models,

    Y . Zhang, Y . Li, L. Cui, D. Cai, L. Liu, T. Fu, X. Huang, E. Zhao, Y . Zhang, Y . Chen,et al., “Siren’s song in the ai ocean: A survey on hal- lucination in large language models,” arXiv preprint arXiv:2309.01219, vol. 2, no. 5, 2023

    2023 arXiv

  9. [11]

    A long way to go: Inves- tigating length correlations in rlhf,

    P. Singhal, T. Goyal, J. Xu, and G. Durrett, “A long way to go: Inves- tigating length correlations in rlhf,” arXiv preprint arXiv:2310.03716 , 2023

    2023 arXiv

  10. [12]

    Human-level control through deep reinforcement learning,

    V . Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski, et al. , “Human-level control through deep reinforcement learning,” nature, vol. 518, no. 7540, pp. 529–533, 2015

    2015

  11. [13]

    Continuous control with deep reinforce- ment learning,

    T. P. Lillicrap, J. J. Hunt, A. Pritzel, N. Heess, T. Erez, Y . Tassa, D. Silver, and D. Wierstra, “Continuous control with deep reinforce- ment learning,” arXiv preprint arXiv:1509.02971 , 2015

    2015 arXiv

  12. [14]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning,

    D. Guo, D. Yang, H. Zhang, J. Song, R. Zhang, R. Xu, Q. Zhu, S. Ma, P. Wang, X. Bi, et al., “Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning,” arXiv preprint arXiv:2501.12948 , 2025

    2025 arXiv

  13. [15]

    Mixcl: Mixed contrastive learning for relation extraction,

    J. Zhang, B. Li, X. Cao, M. Zhang, and W. Zhao, “Mixcl: Mixed contrastive learning for relation extraction,” in Pacific-Asia Conference on Knowledge Discovery and Data Mining . Springer, 2024, pp. 92– 103

    2024

  14. [16]

    Neural path hunter: Reducing hallucination in dialogue systems via path grounding,

    N. Dziri, A. Madotto, O. Za ¨ıane, and A. J. Bose, “Neural path hunter: Reducing hallucination in dialogue systems via path grounding,” arXiv preprint arXiv:2104.08455, 2021

    2021 arXiv

  15. [17]

    Mitigating large language model hallucinations via autonomous knowledge graph- based retrofitting,

    X. Guan, Y . Liu, H. Lin, Y . Lu, B. He, X. Han, and L. Sun, “Mitigating large language model hallucinations via autonomous knowledge graph- based retrofitting,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 16, 2024, pp. 18 126–18 134

    2024

  16. [18]

    A stitch in time saves nine: Detecting and mitigating hallucinations of llms by vali- dating low-confidence generation,

    N. Varshney, W. Yao, H. Zhang, J. Chen, and D. Yu, “A stitch in time saves nine: Detecting and mitigating hallucinations of llms by vali- dating low-confidence generation,” arXiv preprint arXiv:2307.03987 , 2023

    2023 arXiv

  17. [19]

    Reasoning on graphs: Faithful and interpretable large language model reasoning,

    L. Luo, Y .-F. Li, G. Haffari, and S. Pan, “Reasoning on graphs: Faithful and interpretable large language model reasoning,” arXiv preprint arXiv:2310.01061, 2023

    2023 arXiv

  18. [20]

    Open problems and fundamental limitations of reinforcement learning from human feedback,

    S. Casper, X. Davies, C. Shi, T. K. Gilbert, J. Scheurer, J. Rando, R. Freedman, T. Korbak, D. Lindner, P. Freire, et al., “Open problems and fundamental limitations of reinforcement learning from human feedback,” arXiv preprint arXiv:2307.15217 , 2023

    2023 arXiv

  19. [21]

    Selfcheckgpt: Zero-resource black-box hallucination detection for generative large language mod- els,

    P. Manakul, A. Liusie, and M. J. Gales, “Selfcheckgpt: Zero-resource black-box hallucination detection for generative large language mod- els,” arXiv preprint arXiv:2303.08896 , 2023

    2023 arXiv

  20. [22]

    Chain-of-verification reduces hallucination in large language models,

    S. Dhuliawala, M. Komeili, J. Xu, R. Raileanu, X. Li, A. Celikyilmaz, and J. Weston, “Chain-of-verification reduces hallucination in large language models,” arXiv preprint arXiv:2309.11495 , 2023

    2023 arXiv

  21. [23]

    Self-alignment for factuality: Mitigating hallucinations in llms via self-evaluation,

    X. Zhang, B. Peng, Y . Tian, J. Zhou, L. Jin, L. Song, H. Mi, and H. Meng, “Self-alignment for factuality: Mitigating hallucinations in llms via self-evaluation,” arXiv preprint arXiv:2402.09267 , 2024

    2024 arXiv

  22. [24]

    Self-contradictory hallucinations of large language models: Evaluation, detection and mitigation,

    N. M ¨undler, J. He, S. Jenko, and M. Vechev, “Self-contradictory hallucinations of large language models: Evaluation, detection and mitigation,” arXiv preprint arXiv:2305.15852 , 2023

    2023 arXiv

  23. [25]

    Kcts: knowledge-constrained tree search decoding with token-level hallucination detection,

    S. Choi, T. Fang, Z. Wang, and Y . Song, “Kcts: knowledge-constrained tree search decoding with token-level hallucination detection,” arXiv preprint arXiv:2310.09044, 2023

    2023 arXiv

  24. [26]

    Inference- time intervention: Eliciting truthful answers from a language model,

    K. Li, O. Patel, F. Vi ´egas, H. Pfister, and M. Wattenberg, “Inference- time intervention: Eliciting truthful answers from a language model,” 14 Advances in Neural Information Processing Systems , vol. 36, pp. 41 451–41 530, 2023

    2023

  25. [27]

    Decoding by contrasting layers improves factuality in large language models. arxiv 2023,

    Y . Chuang, Y . Xie, H. Luo, Y . Kim, J. Glass, and P. D. He, “Decoding by contrasting layers improves factuality in large language models. arxiv 2023,” arXiv preprint arXiv:2309.03883

    2023 arXiv

  26. [28]

    Self-refine: Iter- ative refinement with self-feedback,

    A. Madaan, N. Tandon, P. Gupta, S. Hallinan, L. Gao, S. Wiegreffe, U. Alon, N. Dziri, S. Prabhumoye, Y . Yang, et al. , “Self-refine: Iter- ative refinement with self-feedback,” Advances in Neural Information Processing Systems, vol. 36, pp. 46 534–46 594, 2023

    2023

  27. [29]

    Can large lan- guage models really improve by self-critiquing their own plans?

    K. Valmeekam, M. Marquez, and S. Kambhampati, “Can large lan- guage models really improve by self-critiquing their own plans?” arXiv preprint arXiv:2310.08118, 2023

    2023 arXiv

  28. [30]

    Large language models lack essential metacognition for reliable medical reasoning,

    M. Griot, C. Hemptinne, J. Vanderdonckt, and D. Yuksel, “Large language models lack essential metacognition for reliable medical reasoning,” Nature communications, vol. 16, no. 1, p. 642, 2025

    2025

  29. [31]

    Large language models cannot self-correct reasoning yet,

    J. Huang, X. Chen, S. Mishra, H. S. Zheng, A. W. Yu, X. Song, and D. Zhou, “Large language models cannot self-correct reasoning yet,” arXiv preprint arXiv:2310.01798 , 2023

    2023 arXiv

  30. [32]

    Llm-based multi-agent rein- forcement learning: Current and future directions,

    C. Sun, S. Huang, and D. Pompili, “Llm-based multi-agent rein- forcement learning: Current and future directions,” arXiv preprint arXiv:2405.11106, 2024

    2024 arXiv

  31. [33]

    Dynamic llm-agent network: An llm-agent collaboration framework with agent team opti- mization,

    Z. Liu, Y . Zhang, P. Li, Y . Liu, and D. Yang, “Dynamic llm-agent network: An llm-agent collaboration framework with agent team opti- mization,” arXiv preprint arXiv:2310.02170 , 2023

    2023 arXiv

  32. [34]

    Leveraging large language models for optimised coordination in textual multi-agent reinforcement learning,

    O. Slumbers, D. H. Mguni, K. Shao, and J. Wang, “Leveraging large language models for optimised coordination in textual multi-agent reinforcement learning,” 2023

    2023

  33. [36]

    Building cooperative embodied agents modularly with large language models,

    H. Zhang, W. Du, J. Shan, Q. Zhou, Y . Du, J. B. Tenenbaum, T. Shu, and C. Gan, “Building cooperative embodied agents modularly with large language models,” arXiv preprint arXiv:2307.02485 , 2023

    2023 arXiv

  34. [37]

    Smart-llm: Smart multi-agent robot task planning using large language models,

    S. S. Kannan, V . L. Venkatesh, and B.-C. Min, “Smart-llm: Smart multi-agent robot task planning using large language models,” in 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024, pp. 12 140–12 147

    2024

  35. [38]

    Roco: Dialectic multi-robot col- laboration with large language models,

    Z. Mandi, S. Jain, and S. Song, “Roco: Dialectic multi-robot col- laboration with large language models,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 286–299

    2024

  36. [39]

    Co-navgpt: Multi-robot cooperative vi- sual semantic navigation using large language models,

    B. Yu, H. Kasaei, and M. Cao, “Co-navgpt: Multi-robot cooperative vi- sual semantic navigation using large language models,” arXiv preprint arXiv:2310.07937, 2023

    2023 arXiv

  37. [40]

    Embodied llm agents learn to cooperate in organized teams,

    X. Guo, K. Huang, J. Liu, W. Fan, N. V ´elez, Q. Wu, H. Wang, T. L. Griffiths, and M. Wang, “Embodied llm agents learn to cooperate in organized teams,” arXiv preprint arXiv:2403.12482 , 2024

    2024 arXiv

  38. [41]

    Multi-agent consensus seeking via large language models,

    H. Chen, W. Ji, L. Xu, and S. Zhao, “Multi-agent consensus seeking via large language models,” arXiv preprint arXiv:2310.20151 , 2023

    2023 arXiv

  39. [42]

    Learning natural language constraints for safe reinforcement learning of language agents,

    J. Chua, C. Wang, and L. Yao, “Learning natural language constraints for safe reinforcement learning of language agents,” arXiv preprint arXiv:2504.03185, 2025

    2025 arXiv

  40. [43]

    Rema: Learning to meta-think for llms with multi-agent reinforcement learning,

    Z. Wan, Y . Li, Y . Song, H. Wang, L. Yang, M. Schmidt, J. Wang, W. Zhang, S. Hu, and Y . Wen, “Rema: Learning to meta-think for llms with multi-agent reinforcement learning,” arXiv preprint arXiv:2503.09501, 2025

    2025 arXiv

  41. [44]

    Theory of mind for multi-agent collaboration via large language models,

    H. Li, Y . Q. Chong, S. Stepputtis, J. Campbell, D. Hughes, M. Lewis, and K. Sycara, “Theory of mind for multi-agent collaboration via large language models,” arXiv preprint arXiv:2310.10701 , 2023

    2023 arXiv

  42. [45]

    Buffer of thoughts: Thought-augmented reasoning with large language models,

    L. Yang, Z. Yu, T. Zhang, S. Cao, M. Xu, W. Zhang, J. E. Gonzalez, and B. Cui, “Buffer of thoughts: Thought-augmented reasoning with large language models,” Advances in Neural Information Processing Systems, vol. 37, pp. 113 519–113 544, 2024

    2024

  43. [46]

    Meta learning for natural language processing: A survey,

    H.-y. Lee, S.-W. Li, and N. T. Vu, “Meta learning for natural language processing: A survey,” arXiv preprint arXiv:2205.01500 , 2022

    2022 arXiv

  44. [48]

    Continual learning of large language models: A comprehensive survey,

    H. Shi, Z. Xu, H. Wang, W. Qin, W. Wang, Y . Wang, Z. Wang, S. Ebrahimi, and H. Wang, “Continual learning of large language models: A comprehensive survey,” arXiv preprint arXiv:2404.16789 , 2024

    2024 arXiv

  45. [50]

    Learn from model beyond fine-tuning: A survey,

    H. Zheng, L. Shen, A. Tang, Y . Luo, H. Hu, B. Du, and D. Tao, “Learn from model beyond fine-tuning: A survey,” arXiv preprint arXiv:2310.08184, 2023

    2023 arXiv

  46. [51]

    Meta rea- soning for large language models,

    P. Gao, A. Xie, S. Mao, W. Wu, Y . Xia, H. Mi, and F. Wei, “Meta rea- soning for large language models,” arXiv preprint arXiv:2406.11698 , 2024

    2024 arXiv

  47. [52]

    Reasoning with large language models, a survey,

    A. Plaat, A. Wong, S. Verberne, J. Broekens, N. van Stein, and T. Back, “Reasoning with large language models, a survey,” arXiv preprint arXiv:2407.11511, 2024

    2024

  48. [55]

    Thinking ma- chines: A survey of llm based reasoning strategies,

    D. Bandyopadhyay, S. Bhattacharjee, and A. Ekbal, “Thinking ma- chines: A survey of llm based reasoning strategies,” arXiv preprint arXiv:2503.10814, 2025

    2025 arXiv

  49. [56]

    Mr-ben: A meta-reasoning benchmark for evaluating system-2 thinking in llms,

    Z. Zeng, Y . Liu, Y . Wan, J. Li, P. Chen, J. Dai, Y . Yao, R. Xu, Z. Qi, W. Zhao, et al., “Mr-ben: A meta-reasoning benchmark for evaluating system-2 thinking in llms,” arXiv preprint arXiv:2406.13975 , 2024

    2024 arXiv

  50. [57]

    Multi-logieval: Towards evaluating multi- step logical reasoning ability of large language models,

    N. Patel, M. Kulkarni, M. Parmar, A. Budhiraja, M. Nakamura, N. Varshney, and C. Baral, “Multi-logieval: Towards evaluating multi- step logical reasoning ability of large language models,” arXiv preprint arXiv:2406.17169, 2024

    2024 arXiv

  51. [58]

    Evaluating the meta-and object-level reasoning of large language models for question answering,

    N. Ferguson, L. Guillou, A. Bundy, and K. Nuamah, “Evaluating the meta-and object-level reasoning of large language models for question answering,” arXiv preprint arXiv:2502.10338 , 2025

    2025 arXiv

  52. [59]

    Malalgoqa: Pedagogical evaluation of counterfactual reasoning in large language models and implications for ai in education,

    S. Sonkar, N. Liu, M. Le, and R. Baraniuk, “Malalgoqa: Pedagogical evaluation of counterfactual reasoning in large language models and implications for ai in education,” in Findings of the Association for Computational Linguistics: EMNLP 2024 , 2024, pp. 15 554–15 567

    2024

  53. [60]

    Metal: To- wards multilingual meta-evaluation,

    R. Hada, V . Gumma, M. Ahmed, K. Bali, and S. Sitaram, “Metal: To- wards multilingual meta-evaluation,” arXiv preprint arXiv:2404.01667, 2024

    2024 arXiv

  54. [61]

    Metaicl: Learn- ing to learn in context,

    S. Min, M. Lewis, L. Zettlemoyer, and H. Hajishirzi, “Metaicl: Learn- ing to learn in context,” arXiv preprint arXiv:2110.15943 , 2021

    2021 arXiv

  55. [62]

    Towards metacognitive clinical reasoning: Benchmarking md-pie against state-of-the-art llms in medical decision-making,

    Y . Esteitieh, S. Mandal, and G. Laliotis, “Towards metacognitive clinical reasoning: Benchmarking md-pie against state-of-the-art llms in medical decision-making,” medRxiv, pp. 2025–01, 2025

    2025

  56. [63]

    Do large language models know what they don’t know?

    Z. Yin, Q. Sun, Q. Guo, J. Wu, X. Qiu, and X. Huang, “Do large language models know what they don’t know?” arXiv preprint arXiv:2305.18153, 2023

    2023 arXiv

  57. [64]

    Language grounded multi- agent reinforcement learning with human-interpretable communica- tion,

    H. Li, H. Nourkhiz Mahjoub, B. Chalaki, V . Tadiparthi, K. Lee, E. Moradi Pari, C. Lewis, and K. Sycara, “Language grounded multi- agent reinforcement learning with human-interpretable communica- tion,” Advances in Neural Information Processing Systems , vol. 37, pp. 87 908–87...

    2024

  58. [65]

    Hypo- thetical minds: Scaffolding theory of mind for multi-agent tasks with large language models,

    L. Cross, V . Xiang, A. Bhatia, D. L. Yamins, and N. Haber, “Hypo- thetical minds: Scaffolding theory of mind for multi-agent tasks with large language models,” arXiv preprint arXiv:2407.07086 , 2024

    2024 arXiv

  59. [66]

    Distilling the knowledge in a neural network,

    G. Hinton, O. Vinyals, and J. Dean, “Distilling the knowledge in a neural network,” arXiv preprint arXiv:1503.02531 , 2015

    2015 arXiv

  60. [67]

    Knowledge distillation: A survey,

    J. Gou, B. Yu, S. J. Maybank, and D. Tao, “Knowledge distillation: A survey,” International Journal of Computer Vision , vol. 129, no. 6, pp. 1789–1819, 2021

    2021

  61. [68]

    Large language models have intrinsic self-correction ability,

    D. Liu, A. Nassereldine, Z. Yang, C. Xu, Y . Hu, J. Li, U. Kumar, C. Lee, R. Qin, Y . Shi, et al. , “Large language models have intrinsic self-correction ability,” arXiv preprint arXiv:2406.15673 , 2024

    2024 arXiv

  62. [69]

    Can rationalization improve robustness?

    H. Chen, J. He, K. Narasimhan, and D. Chen, “Can rationalization improve robustness?” arXiv preprint arXiv:2204.11790 , 2022

    2022 arXiv

  63. [70]

    Measuring compositionality in representation learning,

    J. Andreas, “Measuring compositionality in representation learning,” arXiv preprint arXiv:1902.07181 , 2019

    1902 arXiv

  64. [71]

    Multi-agent reinforcement learning in sequential social dilemmas,

    J. Z. Leibo, V . Zambaldi, M. Lanctot, J. Marecki, and T. Graepel, “Multi-agent reinforcement learning in sequential social dilemmas,” arXiv preprint arXiv:1702.03037 , 2017

    2017 arXiv

  65. [72]

    Towards effective genai multi-agent collaboration: Design and evaluation for enterprise applications,

    R. Shu, N. Das, M. Yuan, M. Sunkara, and Y . Zhang, “Towards effective genai multi-agent collaboration: Design and evaluation for enterprise applications,” arXiv preprint arXiv:2412.05449 , 2024

    2024 arXiv

  66. [73]

    Ai safety via debate,

    G. Irving, P. Christiano, and D. Amodei, “Ai safety via debate,” arXiv preprint arXiv:1805.00899, 2018

    2018 arXiv

  67. [74]

    Adversarial training for high-stakes reliability,

    D. Ziegler, S. Nix, L. Chan, T. Bauman, P. Schmidt-Nielsen, T. Lin, A. Scherlis, N. Nabeshima, B. Weinstein-Raun, D. de Haas, et al. , “Adversarial training for high-stakes reliability,” Advances in neural information processing systems , vol. 35, pp. 9274–9286, 2022. 15

    2022

  68. [76]

    Deep reinforcement learning from human preferences,

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

    2017

  69. [77]

    Training language models to follow instructions with human feedback,

    L. Ouyang, J. Wu, X. Jiang, D. Almeida, C. Wainwright, P. Mishkin, C. Zhang, S. Agarwal, K. Slama, A. Ray, et al. , “Training language models to follow instructions with human feedback,” Advances in neural information processing systems , vol. 35, pp. 27 730–27 744, 2022

    2022

  70. [78]

    Training language models to self-correct via reinforcement learning,

    A. Kumar, V . Zhuang, R. Agarwal, Y . Su, J. D. Co-Reyes, A. Singh, K. Baumli, S. Iqbal, C. Bishop, R. Roelofs, et al., “Training language models to self-correct via reinforcement learning,” arXiv preprint arXiv:2409.12917, 2024

    2024 arXiv

  71. [79]

    A survey of meta-reinforcement learning,

    J. Beck, R. Vuorio, E. Z. Liu, Z. Xiong, L. Zintgraf, C. Finn, and S. Whiteson, “A survey of meta-reinforcement learning,”arXiv preprint arXiv:2301.08028, 2023

    2023 arXiv

  72. [80]

    Eureka: Human- level reward design via coding large language models,

    Y . J. Ma, W. Liang, G. Wang, D.-A. Huang, O. Bastani, D. Ja- yaraman, Y . Zhu, L. Fan, and A. Anandkumar, “Eureka: Human- level reward design via coding large language models,” arXiv preprint arXiv:2310.12931, 2023

    2023 arXiv

  73. [81]

    Meta-rewarding language models: Self-improving alignment with llm-as-a-meta-judge,

    T. Wu, W. Yuan, O. Golovneva, J. Xu, Y . Tian, J. Jiao, J. Weston, and S. Sukhbaatar, “Meta-rewarding language models: Self-improving alignment with llm-as-a-meta-judge,” arXiv preprint arXiv:2407.19594, 2024

    2024 arXiv

  74. [82]

    Learning to summarize with human feedback,

    N. Stiennon, L. Ouyang, J. Wu, D. Ziegler, R. Lowe, C. V oss, A. Radford, D. Amodei, and P. F. Christiano, “Learning to summarize with human feedback,” Advances in neural information processing systems, vol. 33, pp. 3008–3021, 2020

    2020

  75. [83]

    Confidence matters: Revisiting intrinsic self-correction capabilities of large language models,

    L. Li, Z. Chen, G. Chen, Y . Zhang, Y . Su, E. Xing, and K. Zhang, “Confidence matters: Revisiting intrinsic self-correction capabilities of large language models,” arXiv preprint arXiv:2402.12563 , 2024

    2024 arXiv

  76. [84]

    Curiosity- driven reinforcement learning from human feedback,

    H. Sun, Y . Chai, S. Wang, Y . Sun, H. Wu, and H. Wang, “Curiosity- driven reinforcement learning from human feedback,” arXiv preprint arXiv:2501.11463, 2025

    2025 arXiv

  77. [85]

    Online intrinsic rewards for decision making agents from large language model feedback,

    Q. Zheng, M. Henaff, A. Zhang, A. Grover, and B. Amos, “Online intrinsic rewards for decision making agents from large language model feedback,” arXiv preprint arXiv:2410.23022 , 2024

    2024

  78. [86]

    Deepseekmath: Pushing the limits of mathematical reasoning in open language models,

    Z. Shao, P. Wang, Q. Zhu, R. Xu, J. Song, X. Bi, H. Zhang, M. Zhang, Y . Li, Y . Wu,et al., “Deepseekmath: Pushing the limits of mathematical reasoning in open language models,” arXiv preprint arXiv:2402.03300, 2024

    2024 arXiv

  79. [87]

    Grandmaster level in starcraft ii using multi-agent reinforcement learning,

    O. Vinyals, I. Babuschkin, W. M. Czarnecki, M. Mathieu, A. Dudzik, J. Chung, D. H. Choi, R. Powell, T. Ewalds, P. Georgiev, et al. , “Grandmaster level in starcraft ii using multi-agent reinforcement learning,” nature, vol. 575, no. 7782, pp. 350–354, 2019

    2019

  80. [88]

    Human-level play in the game of diplomacy by combining language models with strategic reasoning,

    M. F. A. R. D. T. (FAIR)†, A. Bakhtin, N. Brown, E. Dinan, G. Farina, C. Flaherty, D. Fried, A. Goff, J. Gray, H. Hu, et al. , “Human-level play in the game of diplomacy by combining language models with strategic reasoning,” Science, vol. 378, no. 6624, pp. 1067–1074, 2022

    2022

  81. [89]

    Shall we team up: Exploring spontaneous co- operation of competing llm agents,

    Z. Wu, R. Peng, S. Zheng, Q. Liu, X. Han, B. I. Kwon, M. Onizuka, S. Tang, and C. Xiao, “Shall we team up: Exploring spontaneous co- operation of competing llm agents,” arXiv preprint arXiv:2402.12327 , 2024

    2024 arXiv

  82. [90]

    Self- playing adversarial language game enhances llm reasoning,

    P. Cheng, T. Hu, H. Xu, Z. Zhang, Y . Dai, L. Han, X. Li, et al., “Self- playing adversarial language game enhances llm reasoning,” Advances in Neural Information Processing Systems , vol. 37, pp. 126 515– 126 543, 2024

    2024

  83. [91]

    Logicattack: Adversarial attacks for evaluating logical consistency of natural language inference,

    M. Nakamura, S. Mashetty, M. Parmar, N. Varshney, and C. Baral, “Logicattack: Adversarial attacks for evaluating logical consistency of natural language inference,” in Findings of the Association for Computational Linguistics: EMNLP 2023 , 2023, pp. 13 322–13 334

    2023

  84. [92]

    Model-agnostic meta-learning for fast adaptation of deep networks,

    C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” in International conference on machine learning. PMLR, 2017, pp. 1126–1135

    2017

  85. [93]

    Meta-learning for few-shot natural language processing: A survey,

    W. Yin, “Meta-learning for few-shot natural language processing: A survey,” arXiv preprint arXiv:2007.09604 , 2020

    2007 arXiv

  86. [94]

    Meta in- context learning makes large language models better zero and few-shot relation extractors,

    G. Li, P. Wang, J. Liu, Y . Guo, K. Ji, Z. Shang, and Z. Xu, “Meta in- context learning makes large language models better zero and few-shot relation extractors,” arXiv preprint arXiv:2404.17807 , 2024

    2024 arXiv

  87. [95]

    Improving consistency in large language models through chain of guidance,

    H. Raj, V . Gupta, D. Rosati, and S. Majumdar, “Improving consistency in large language models through chain of guidance,” arXiv preprint arXiv:2502.15924, 2025

    2025 arXiv

  88. [96]

    Training verifiers to solve math word problems,

    K. Cobbe, V . Kosaraju, M. Bavarian, M. Chen, H. Jun, L. Kaiser, M. Plappert, J. Tworek, J. Hilton, R. Nakano, et al., “Training verifiers to solve math word problems,” arXiv preprint arXiv:2110.14168, 2021

    2021 arXiv

  89. [97]

    Red teaming language models for contradictory dialogues,

    X. Wen, B. Li, T. Huang, and M. Chen, “Red teaming language models for contradictory dialogues,” arXiv e-prints, pp. arXiv–2405, 2024

    2024

  90. [98]

    Openai o1 system card,

    A. Jaech, A. Kalai, A. Lerer, A. Richardson, A. El-Kishky, A. Low, A. Helyar, A. Madry, A. Beutel, A. Carney, et al., “Openai o1 system card,” arXiv preprint arXiv:2412.16720 , 2024

    2024 arXiv

  91. [99]

    Did aristotle use a laptop? a question answering benchmark with implicit reasoning strategies,

    M. Geva, D. Khashabi, E. Segal, T. Khot, D. Roth, and J. Berant, “Did aristotle use a laptop? a question answering benchmark with implicit reasoning strategies,” Transactions of the Association for Computational Linguistics, vol. 9, pp. 346–361, 2021

    2021

  92. [100]

    Can llms beat humans in debating? a dynamic multi-agent framework for competitive debate,

    Y . Zhang, X. Yang, S. Feng, D. Wang, Y . Zhang, and K. Song, “Can llms beat humans in debating? a dynamic multi-agent framework for competitive debate,” arXiv preprint arXiv:2408.04472 , 2024

    2024 arXiv

  93. [101]

    Scalable evaluation of multi-agent reinforcement learning with melting pot,

    J. Z. Leibo, E. A. Due ˜nez-Guzman, A. Vezhnevets, J. P. Agapiou, P. Sunehag, R. Koster, J. Matyas, C. Beattie, I. Mordatch, and T. Grae- pel, “Scalable evaluation of multi-agent reinforcement learning with melting pot,” in International conference on machine learning. PMLR, 2...

    2021

  94. [102]

    Beyond the imitation game: Quantifying and extrapolating the capabilities of language models,

    A. Srivastava, A. Rastogi, A. Rao, A. A. M. Shoeb, A. Abid, A. Fisch, A. R. Brown, A. Santoro, A. Gupta, A. Garriga-Alonso, et al., “Beyond the imitation game: Quantifying and extrapolating the capabilities of language models,” arXiv preprint arXiv:2206.04615 , 2022

    2022 arXiv

  95. [103]

    Sciinstruct: A self-reflective instruction an- notated dataset for training scientific language models,

    D. Zhang, Z. Hu, S. Zhoubian, Z. Du, K. Yang, Z. Wang, Y . Yue, Y . Dong, and J. Tang, “Sciinstruct: A self-reflective instruction an- notated dataset for training scientific language models,” Advances in Neural Information Processing Systems, vol. 37, pp. 1443–1473, 2024

    2024

  96. [104]

    Debateqa: Evalu- ating question answering on debatable knowledge,

    R. Xu, X. Qi, Z. Qi, W. Xu, and Z. Guo, “Debateqa: Evalu- ating question answering on debatable knowledge,” arXiv preprint arXiv:2408.01419, 2024

    2024 arXiv

  97. [105]

    A survey on multi-agent deep reinforcement learning: from the perspective of challenges and applications,

    W. Du and S. Ding, “A survey on multi-agent deep reinforcement learning: from the perspective of challenges and applications,”Artificial Intelligence Review, vol. 54, no. 5, pp. 3215–3238, 2021

    2021

  98. [106]

    Dealing with non-stationarity in multi-agent deep reinforcement learning,

    G. Papoudakis, F. Christianos, A. Rahman, and S. V . Albrecht, “Dealing with non-stationarity in multi-agent deep reinforcement learning,” arXiv preprint arXiv:1906.04737 , 2019

    1906 arXiv

  99. [107]

    Games of gans: Game- theoretical models for generative adversarial networks,

    M. Mohebbi Moghaddam, B. Boroomand, M. Jalali, A. Zareian, A. Daeijavad, M. H. Manshaei, and M. Krunz, “Games of gans: Game- theoretical models for generative adversarial networks,” Artificial In- telligence Review, vol. 56, no. 9, pp. 9771–9807, 2023

    2023

  100. [108]

    Sycophancy to subterfuge: Investigating reward-tampering in large language models,

    C. Denison, M. MacDiarmid, F. Barez, D. Duvenaud, S. Kravec, S. Marks, N. Schiefer, R. Soklaski, A. Tamkin, J. Kaplan, et al. , “Sycophancy to subterfuge: Investigating reward-tampering in large language models,” arXiv preprint arXiv:2406.10162 , 2024

    2024 arXiv

  101. [109]

    Diverse auto-curriculum is critical for successful real-world multiagent learning systems,

    Y . Yang, J. Luo, Y . Wen, O. Slumbers, D. Graves, H. B. Ammar, J. Wang, and M. E. Taylor, “Diverse auto-curriculum is critical for successful real-world multiagent learning systems,” arXiv preprint arXiv:2102.07659, 2021

    2021 arXiv

  102. [110]

    Multi-agent systems: overview of a new paradigm for distributed systems,

    H. F. Ahmad, “Multi-agent systems: overview of a new paradigm for distributed systems,” in 7th IEEE International Symposium on High Assurance Systems Engineering, 2002. Proceedings. IEEE, 2002, pp. 101–107

    2002

  103. [111]

    Pyramidtabnet: Transformer-based table recognition in image-based documents,

    M. Umer, M. A. Mohsin, A. Ul-Hasan, and F. Shafait, “Pyramidtabnet: Transformer-based table recognition in image-based documents,” in Document Analysis and Recognition - ICDAR 2023, G. A. Fink, R. Jain, K. Kise, and R. Zanibbi, Eds. Cham: Springer Nature Switzerland, 2023, pp. 420–437

    2023

  104. [112]

    Multi-agent systems: A survey,

    A. Dorri, S. S. Kanhere, and R. Jurdak, “Multi-agent systems: A survey,” Ieee Access, vol. 6, pp. 28 573–28 593, 2018

    2018

  105. [113]

    Distributed task allocation in multi-agent systems,

    N. Krothapalli and A. V . Deshmukh, “Distributed task allocation in multi-agent systems,” in Proceedings of the Institute of Industrial Engineers Annual Conference . Citeseer, 2002

    2002

  106. [114]

    Taxonomy of risks posed by language models,

    L. Weidinger, J. Uesato, M. Rauh, C. Griffin, P.-S. Huang, J. Mellor, A. Glaese, M. Cheng, B. Balle, A. Kasirzadeh, et al. , “Taxonomy of risks posed by language models,” in Proceedings of the 2022 ACM conference on fairness, accountability, and transparency , 2022, pp. 214–229

    2022

  107. [115]

    Model cards for model reporting,

    M. Mitchell, S. Wu, A. Zaldivar, P. Barnes, L. Vasserman, B. Hutchin- son, E. Spitzer, I. D. Raji, and T. Gebru, “Model cards for model reporting,” in Proceedings of the conference on fairness, accountability, and transparency, 2019, pp. 220–229

    2019

  108. [116]

    Neuroscience-inspired artificial intelligence,

    D. Hassabis, D. Kumaran, C. Summerfield, and M. Botvinick, “Neuroscience-inspired artificial intelligence,” Neuron, vol. 95, no. 2, pp. 245–258, 2017

    2017

  109. [117]

    The consciousness prior,

    Y . Bengio, “The consciousness prior,” arXiv preprint arXiv:1709.08568, 2017

    2017 arXiv

  110. [118]

    Neurosymbolic reinforcement learning and planning: A survey,

    K. Acharya, W. Raza, C. Dourado, A. Velasquez, and H. H. Song, “Neurosymbolic reinforcement learning and planning: A survey,” IEEE 16 Transactions on Artificial Intelligence , vol. 5, no. 5, pp. 1939–1953, 2023

    1939

  111. [119]

    Neuro- symbolic artificial intelligence: a survey,

    B. P. Bhuyan, A. Ramdane-Cherif, R. Tomar, and T. Singh, “Neuro- symbolic artificial intelligence: a survey,” Neural Computing and Applications, vol. 36, no. 21, pp. 12 809–12 844, 2024

    2024

  112. [120]

    Peorl: Integrating symbolic planning and hierarchical reinforcement learning for robust decision- making,

    F. Yang, D. Lyu, B. Liu, and S. Gustafson, “Peorl: Integrating symbolic planning and hierarchical reinforcement learning for robust decision- making,” arXiv preprint arXiv:1804.07779 , 2018

    2018 arXiv

  113. [121]

    The programmer’s assistant: Conversational interaction with a large language model for software development,

    S. I. Ross, F. Martinez, S. Houde, M. Muller, and J. D. Weisz, “The programmer’s assistant: Conversational interaction with a large language model for software development,” in Proceedings of the 28th International Conference on Intelligent User Interfaces, 2023, pp. 491– 514

    2023

  114. [122]

    Onrl-rag: Real-time personalized mental health dialogue system,

    A. Bilal, B. Lin, and M. Zaeifi, “Onrl-rag: Real-time personalized mental health dialogue system,” arXiv preprint arXiv:2504.02894 , 2025

    2025 arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.