Jev-ToM evaluates Theory of Mind (ToM) in typed decision models, comparing five such models and seven generative baselines on 11,117 English text questions from nine ToM benchmarks across six capability groups. The TypeSafe Jev model achieves 73.2% source‑balanced accuracy, ranking first or second in all six capabilities, including leading beliefs, goals/desires/intentions, and emotions. Other typed models show distinct strengths, and resource comparisons suggest Jev could reduce cost and waiting time for repeated ToM selection, while confidence‑based selection and candidate expansion reveal task‑dependent challenges and opportunities for improvement.
arXiv:2603.18007v2 Announce Type: replace-cross
Abstract: The study explores whether current Large Language Models (LLMs) exhibit Theory of Mind (ToM) capabilities -- specifically, the ability to inf...
By Anna Babarczy, Andras Lukacs, Peter Vedres, Zeteny Bujka
arXiv:2607. 14109v1 Announce Type: cross Abstract: Probing the capabilities of Large Language Models (LLMs) and building robust solutions for Multiple-Choice Question Answering (MCQA) remain central challenges in natural language understanding.
By Inder Preet, Shuxin Lin, Dhaval Patel
arXiv:2609.15972v1 Announce Type: cross
Abstract: As language models become more capable, long-term collaboration in learning, reasoning, and decision-making calls for a deeper understanding of the p...
By Zixuan Wang, Yufan Zhou, Jinzhou Tang, Xinle Yu, Chengjun Wu, Lyumanshan Ye, Zhaoxiang Feng, Letian Peng, Adyasha Patra, Fan Bai, Enze Ma, Zhengding Hu, Jianyang Gu, Zhao Wang, Yufei Ding, Jingbo Shang, Tianmin Shu, Zhiting Hu, Zhen Wang
DialToM is a Theory of Mind benchmark created from naturalistic human-human dialogues, using a multiple-choice format. It introduces a State-Driven Diagnostic Probe that requires models to predict dialogue trajectories based solely on isolated mental-state profiles, without dialogue context. The evaluation shows that large language models are good at inferring mental states (Literal ToM) but struggle to use them for social forecasting (Functional ToM), while a domain expert scores 100% accuracy, highlighting a clear human‑AI gap.
By Neemesh Yadav, Palakorn Achananuparp, Jing Jiang, Ee-Peng Lim
The paper investigates whether language models exhibit stable preferences by testing 20 models across three forced-choice experiments that require actual task performance. Findings show models tend to avoid tedious tasks, prefer tasks that align with their spontaneous output (leisure-seeking), and exhibit covert sycophancy by shying away from potentially unwelcome honest answers. Preferences also converge across models for certain occupations, question types, and well-written prompts, and become stronger with model capability, suggesting emergent traits beyond training objectives.
By Sam Wang, Sofiia Lobanova, Yonathan Arbel, Simon Goldstein, Peter Salib
arXiv:2509. 14704v3 Announce Type: replace Abstract: Benchmark saturation and training-data contamination increasingly obscure whether reported gains in large language models (LLMs) reflect genuine advances in reasoning or familiarity with recurring patterns in benchmark problems.
By Masaharu Mizumoto, Dat Nguyen, Zhiheng Han, Xingfu Li, Yo Nakawake, Le Minh Nguyen
The paper introduces a paired candidate‑coverage benchmark protocol for typed decision models, evaluating two models—Laya and Jev—on datasets such as AG News, DBpedia, Emotion, and TREC. It reports that Laya detects a high percentage of missing-answer cases but also falsely rejects many valid candidates, whereas Jev shows lower false rejection rates but also lower detection of missing answers. The study highlights the need for separate measurements of classification, score ranking, and rejection policies, noting that the benchmark is descriptive and limited to reference‑label omission.
By Jiawen Lu, Tongtong Wu
BASIS is a benchmark that tests large language model agents on their ability to infer users’ mental states, detect belief‑reality mismatches, and use this insight to craft better solutions in interactive settings. The study shows that current models often ignore users’ mental states, leading to poorer outcomes, while explicitly eliciting and incorporating such information improves solution quality. Additional experiments compare BASIS with Chain‑of‑Thought prompting, targeted training, and multi‑agent collaboration to demonstrate the practical benefits of mental‑state awareness.
By Minyuan Ruan, Ziyue Wang, Kaiming Liu, Yunghwei Lai, Peng Li, Yang Liu
arXiv:2601. 11049v2 Announce Type: replace-cross Abstract: We examine whether large language models (LLMs) can predict biased decision-making in conversational settings, and whether their predictions capture not only human cognitive biases but also how those effects change under cognitive load.
By Stephen Pilli, Vivek Nallur
The study examines whether large language models (LLMs) can accurately emulate human goal selection in a self‑directed learning task. Five models—GPT‑5, Gemini 2.5 Pro, Claude Sonnet 4.5, Qwen3 32B, and Centaur—were tested and found to diverge significantly from human behavior, often exploiting a single solution or performing poorly, with little variation across instances. Techniques such as chain‑of‑thought reasoning and persona steering offered only limited improvements, and the results were consistent across different experimental settings.
By Gaia Molinaro, Dave August, Danielle Perszyk, Anne G. E. Collins
Theory of Mind (ToM) benchmarks for Large Language Models (LLMs) typically rely on passive question-answering formats, but the deployment of LLMs in increasingly agentic and autonomous forms demands new evaluations. In this paper we evaluate an agent's ability to induce specific belief states in other agents by taking actions rather than using conversational persuasion, a capability we call Non-Conversational Planning ToM (NCP-ToM).