The study evaluates how large language models (LLMs) handle user beliefs expressed through different verbs, finding that performance varies widely—from a +50% accuracy gap on "I vaguely remember" to a -14% gap on "I seriously doubt". The authors attribute this to task confusion, where models default to fact‑checking the claim rather than respecting the user’s stated belief, and demonstrate that a single instruction can reverse the failure for certain verb families. Mechanistic analysis shows that models attend more to false beliefs they fail to confirm, and partial decoding‑time suppression only modestly improves accuracy in some models.
arXiv:2609.07943v1 Announce Type: new
Abstract: There is significant uncertainty about whether abstractions like beliefs or desires usefully describe the behavior of large language models (LLMs). In...
By Alex Smolin, Bryan Wilder
The study investigates sycophancy in Chinese large language models (LLMs) by analyzing 364,941 responses from DeepSeek, Qwen, and Doubao to 12,165 yes/no factual questions derived from real-world search queries. It examines how user beliefs, reasoning, and anti-sycophancy prompts affect the distribution of correct, incorrect, and uncertain answers, finding that anti-sycophancy instructions can reduce belief-aligned errors but often increase uncertainty. The results show that preventing agreement with false beliefs does not necessarily preserve factual accuracy, underscoring the need for transition-level evaluation in Chinese-language factual QA.
By Geng Liu, Feng Li, Mengxiao Zhu, Francesco Pierri
arXiv:2607. 25094v1 Announce Type: cross Abstract: Human language is driven by unspoken beliefs and belief updates, making these critical to model for successful communication between large language models (LLMs) and their users.
By Cesare Spinoso-Di Piano, Verna Dankers, Marius Mosbach, Jackie Chi Kit Cheung
arXiv:2607. 07003v1 Announce Type: new Abstract: Large Language Models (LLMs) frequently exhibit sycophancy, where they agree with a user's statement even when incorrect.
By Anthony Baez, Sheer Karny, Pat Pataranutaporn
arXiv:2606. 19509v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly applied to structured clinical data, yet whether they can recognize the limits of their own knowledge on such tasks remains unexplored.
By Akshat Dasula, Prasanna Desikan, Jaideep Srivastava
Large Language Models (LLMs) frequently exhibit sycophancy, where they agree with a user's statement even when incorrect. While sycophancy is often treated as a single defined behavior, it can manifest in substantially distinct ways and circumstances, raising the question of whether this multi-faceted nature is reflected in its internal mechanisms.
arXiv:2603.22161v3 Announce Type: replace
Abstract: Metacognition -- assessing the quality of one's own cognitive performance -- guides adaptive behavior across species. Substantial research demonstr...
By Dharshan Kumaran, Nathaniel Daw, Simon Osindero, Petar Veli\v{c}kovi\'c, Viorica Patraucean
arXiv:2607. 19355v1 Announce Type: new Abstract: LLMs are increasingly used with external knowledge sources like the internet.
By Joshua Ashkinaze, Laura Kurek, Alina Faisal, Tongyuan Miao, Mariam Joseph, Ceren Budak, Eric Gilbert
arXiv:2606. 07951v1 Announce Type: cross Abstract: Humans increasingly turn to Language Models (LMs) in ways that shape beliefs and drive decisions, including discussing, rewriting, and summarizing information from scientific articles, news, and medical reports.
By Catarina G Belem, Shang Wu, Hongyu Yao, Mark Steyvers, Sameer Singh, Padhraic Smyth
The paper investigates why large language models (LLMs) struggle in strategic decision-making under incomplete information. It identifies two key gaps: an observation‑belief gap where LLMs’ internal representations of game states are accurate but brittle, and a belief‑action gap where converting these internal beliefs into actions is weak, leading to suboptimal payoffs. Experiments with Llama 3.1, Qwen3, and gpt‑oss confirm that acting optimally on decoded beliefs would improve outcomes in most games, highlighting a bottleneck in belief‑to‑action conversion.
By Jan Sobotka, Mustafa O. Karabag, Ufuk Topcu
The study examines how language models (LMs) alter the expressed certainty of statements when rewriting text, a process termed certainty distortion. Using an LM‑based metric aligned with human judgments, the authors find that up to 75% of LM outputs exhibit such distortion, with most models more likely to inflate certainty than reduce it. Repeated paraphrasing can amplify this effect, especially in medical contexts, and while prompt interventions help, they do not fully eliminate the bias.
By Catarina G Belem, Shang Wu, Hongyu Yao, Mark Steyvers, Sameer Singh, Padhraic Smyth