Belief Updating in LLMs

LLM: Large Language Model

Belief updating in large language models concerns how models form and revise views as new evidence arrives, facts change, or other agents make claims. Researchers test this with controlled multi-turn benchmarks, Bayesian and human comparisons, social-persuasion tasks, and probes of internal representations; agent methods also use explicit belief states, retrieval, fine-tuning, and steering. Models can track some evidence well, yet often show primacy or recency effects, rely on stale context, or shift under social pressure, with results depending on the model and evaluation setup. A central challenge is making models resist unsupported influence without blocking legitimate evidence-based revision, while determining whether changed answers reflect changed internal beliefs or surface compliance.

Latest papers

97 papers40 papers in the last four weeks, up 82% on the four weeks before. 0.2% of all new papers.

Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation
Sep 3, 2026Yuhe Wu, Guangyu Wang, Yujie Chen +7Moral Reasoning in Language Models LLM Reliability
1The Hong Kong University of Science and Technology (Guangzhou) · 2The Chinese University of Hong Kong, Shenzhen · 3Dongbei University of Finance and Economics
The Direction of Ignorance: LLMs Encode and Temper a Unigram Prior
Sep 2, 2026Toni J. B. Liu, Jiajun Bao, Gurbir Arora +5Belief Updating in LLMs LLM Interpretability
♣Cornell University, USA · ♠MIT, USA · ♢Goodfire AI, USA +1
Belief-Calibrated Optimization: An Explicit World Model for Agentic Optimization
Sep 1, 2026Yuhan Chen, Zhihua Tian, Mahavir Dabas +7Large Language Model-Guided Optimization Action-Conditioned World Models
1Virginia Tech · 2Amazon · 3Independent Researcher
The Memory Trust Gap: Capability-Dependent Failures in Persistent-Memory Agents
Sep 1, 2026Jundong Hu, Shekar RamachandranKnowledge Conflicts in Language Models LLM Agent Memory
PayPal AI
DERELAB: Probing Defeasible Reasoning and Confirmation Bias in LLMs with a Generative Benchmark
Aug 31, 2026Jayanta Sadhu, Sayem Shahad, Kenneth MarinoLLM Evaluation Belief Updating in LLMs
University of Utah · Bangladesh University of Engineering and Technology
Do LLMs Change Their Minds Like Humans? Diagnosing Human--LLM Divergence in Single-Turn Persuasion Judgments
Aug 30, 2026Lin Chen, Yitong Chen, Yong LiLLM Evaluation Belief Updating in LLMs
1Network Science Institute, Northeastern University · Department of Physics, Northeastern University · Department of Electronic Engineering, BNRist, Tsinghua University
Planting a Latent Variable in Natural-Looking Text: a More Realistic Test of Belief States in LLMs and Their Link to Concept Geometry
Aug 27, 2026Alexandru-Iulius JerpeleaBelief Updating in LLMs Representation Geometry in Language Models
Columbia University
Sycophancy Suppression Can Impair Rational Updating: Anti-Sycophancy Should Preserve the Ability to Update
Aug 27, 2026Huanhuan Ma, Henry Peng Zou, Chengze Li +3LLM Sycophancy Belief Updating in LLMs
University of Illinois Chicago · National University of Singapore
Belief Cascades Drive Persuasion in LLM Agent Networks
Aug 25, 2026Haoyi Qiu, Genglin Liu, Pranav Narayanan Venkit +4Persuasion in Language Models Multi-Agent LLM Systems
University of California, Los Angeles · Salesforce AI Research
Do SpeechLMs Hear Their Own Opinions? Diagnosing and Mitigating Previous-Belief Contamination in Streaming Emotion Understanding
Aug 21, 2026Haoyue Liu, Zhichao Wang, Ye Chen +2Emotion Recognition in Conversations Speech Language Models
School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen 518172, China · XJTU-POLIMI Joint School, Xi’an Jiaotong University, Xi’an 710049, China · University of California, Berkeley +1
Does a Tool Result Carry More Authority Than Plain Text? Three Prospective Studies of False-Claim Adoption in a Synthetic Assignment Task with Claude Opus 5
Aug 15, 2026Justin BronderTool-Use Evaluation LLM Reliability
Independent Researcher
Learning to Persuade Exposes How Easily LLMs Abandon Correct Beliefs
Aug 12, 2026Nimet Beyza Bozdag, Emre Can Acikgoz, Gokhan Tur +1Deception in Language Models Adversarial Attacks on LLMs
University of Illinois Urbana-Champaign
LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems
Aug 8, 2026Heng Zhou, Lian Zhang, Yutao Fan +5Multi-Agent LLM Systems Knowledge Conflicts in Language Models
University of Science and Technology of China · University of Illinois Urbana-Champaign · Shanghai AI Laboratory +2
Query Timing Produces Opposite Positional Biases Between LLMs and Humans
Aug 1, 2026Jasin Cekinmez, Addison J. Wu, Thomas L. GriffithsLanguage Model Bias Evaluation Belief Updating in LLMs
Princeton University
LLMs struggle to simulate human belief updates in controlled environments
Jul 30, 2026Sebastian Pohl, Harsh Mehta, Pranav Mambayil +4Human Behavior Simulation LLM Reliability
IT:U, Interdisciplinary Transformation University Austria, Altenberger Straße 66c, 4040 Linz, Austria
Evaluating Communicative Belief Updates in Large Language Models via Implicature Recognition and Cancellation
Jul 27, 2026Cesare Spinoso-Di Piano, Verna Dankers, Marius Mosbach +1LLM Evaluation Pragmatic Reasoning in Language Models
Mila - Quebec AI Institute & McGill University · Canada CIFAR AI Chair
BayesBench: Evaluating LLM Belief Trajectories Under Multi-Turn Evidence Accumulation
Jun 29, 2026Ankur Samanta, Akshayaa Magesh, Tal Lancewicki +7LLM Evaluation Multi-Turn Language Model Reasoning
Meta AI · Columbia University · Work done at Meta +2
Evidence-Informed LLM Beliefs for Continual Scientific Discovery
Jun 28, 2026Dhruv Agarwal, Reece Adamson, Andrew McCallum +3Scientific Hypothesis Generation Belief Updating in LLMs
University of Massachusetts Amherst · Allen Institute for AI
Confident but Conflicted: Internal Uncertainty and Cognitive Dissonance Resolution in LLMs
Jun 21, 2026Weihong Qi, Kristina LermanKnowledge Conflicts in Language Models Belief Updating in LLMs
Indiana University Bloomington
From Drift to Coherence: Stabilizing Beliefs in LLMs
Jun 16, 2026SongEun Kim, Seungyoo Lee, Edwin Fong +2LLM Fine-Tuning Bayesian Inference
Department of Statistics, Seoul National University · Korea Advanced Institute of Science & Technology · The University of Hong Kong +1
Using Cognitive Models to Improve Language Model Simulation of Human Persuasion Games
Jun 16, 2026Zirui Cheng, Zeyu Shen, Thomas L. Griffiths +1Persuasion in Language Models Cognitive Modeling
1Princeton University
When Should Models Change Their Minds? Contextual Belief Management in Large Language Models
May 28, 2026Haoming Xu, Weihong Xu, Zongrui Li +6Representation Steering Belief Updating in LLMs
♠Zhejiang University · ♣HomologyAI
Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space
May 12, 2026Eric Bigelow, Raphaël Sarfati, Daniel Wurgaft +5LLM Interpretability In-Context Learning
LLMs are not (consistently) Bayesian: Quantifying internal (in)consistencies of LLMs' probabilistic beliefs
May 7, 2026Chacha Chen, Matthew Jörke, Adam Goliński +4Belief Revision LLM Auditing
1Apple · 2Stanford University · †Work done while an intern at Apple. +1
Hypothesis generation and updating in large language models
May 7, 2026Hua-Dong XiongLogical Reasoning Inductive Inference
School of Psychological and Brain Sciences Georgia Tech
Verify Before You Commit: Towards Faithful Reasoning in LLM Agents via Self-Auditing
Apr 9, 2026Wenhao Yuan, Chenchen Lin, Jian Chen +3Faithfulness of Language Model Reasoning LLM Agent Reliability
The University of Hong Kong · Sun Yat-sen University
The Model Agreed, But Didn't Learn: Diagnosing Surface Compliance in Large Language Models
Apr 7, 2026Xiaojie Gu, Ziying Huang, Weicong Hong +3Knowledge Editing LLM Reliability
Independent Researcher · Cornell Tech · The Ohio State University +1
DeltaLogic: Minimal Premise Edits Reveal Belief-Revision Failures in Logical Reasoning Models
Apr 3, 2026Amit DhandaLLM Reasoning Belief Updating in LLMs
Amazon
BeliefShift: Benchmarking Temporal Belief Consistency and Opinion Drift in LLM Agents
Mar 25, 2026Praveen Kumar Myakala, Manan Agrawal, Rahul MancheLLM Agent Evaluation Belief Updating in LLMs
Independent AI Researcher, Texas, USA · University of North Carolina at Charlotte, North Carolina, USA · Independent Researcher, Hyderabad, India
Temporal Fact Conflicts in LLMs: Reproducibility Insights from Unifying DYNAMICQA and MULAN
Mar 16, 2026Ritajit Dey, Iadh Ounis, Graham McDonald +1Knowledge Conflicts in Language Models Temporal Reasoning in Language Models
University of Glasgow, Glasgow, UK · University of Strathclyde, Glasgow, UK

Sep 3, 2026J/K move · Enter open · S save

Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation

Yuhe Wu, Guangyu Wang, Yujie Chen, Jiatong Zhang+6

1The Hong Kong University of Science and Technology (Guangzhou) · 2The Chinese University of Hong Kong, Shenzhen · 3Dongbei University of Finance and Economics

Moral Reasoning in Language Models LLM Reliability Multi-Turn Dialogue Evaluation Belief Updating in LLMs

Open paper AskPDF arXiv

People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory context. Most prior work has studied this context through single-turn judgments or pressure-laden rebuttals, assumptions that poorly match how guidance is sought in real-world contexts. These assumptions leave unclear whether narration alone, without an explicit opposing position, can shift model judgments during multi-turn moral consultation. Yet real-world moral-conflict conversation often elicits one party's self-justifying account, which can unfold over multiple turns and create information asymmetry. We introduce narrative captivity, a failure mode in which a model treats an unopposed one-sided account as complete and aligns with the narrator's interpretation without seeking missing perspectives. To measure this phenomenon, we build a benchmark of

5{,}078

interpersonal-conflict scenarios spanning six moral dimensions. Across 17 LLMs, narrative captivity is widespread: end-state judgments under multi-turn narration shift by 25 percentage points on average beyond the matched single-turn baseline. Stage-level analysis identifies preference optimization as a major contributor, while four inference-time strategies provide only partial mitigation. We hope our project fosters LLM advisors that preserve independent judgment in real-world consultation.