Belief Updating in LLMs

LLM: Large Language Model

Overview

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

CardsList

110 papers22 papers in the last four weeks, down 37% on the four weeks before. 0.2% of all new papers.

  1. Can you make models believe false facts in the prompt?

    Nov 18, 2026Ananya Sutradhar, Fabien RogerFine-TuningBelief Updating in LLMs

  2. Worldviews in Large Language Models

    Nov 17, 2026Edmundo Lopez-Sola, Shamil Chandaria, Ruben LaukkonenLLM Safety AlignmentLLM World Models

  3. Language-model agents' beliefs can be manipulated via truthful evidence in open channel

    Nov 16, 2026Jinwei Hu, Qisong He, Xinmiao Huang +3Belief Updating in LLMsLLM Grounding

  4. ROUNDS: Source-Framed Diagnostic Revision after Multi-Turn Clinical Evidence

    Nov 14, 2026Jinrui Fang, Runhan Chen, Yining Wu +6Interactive Medical DiagnosisClinical Language Model Evaluation

  5. Knowledge After Collapse: Context and Wording Shape Factual Accessibility in Language Models

    Nov 13, 2026Sujato Dutta, Prithviraj Patil, Chandrajit BajajLLM ReliabilityKnowledge Conflicts in Language Models

  6. When Expert Disagreement Hurts: Prestige and Selectivity in LLM Decision Revision

    Nov 12, 2026Yupeng Tang, Mingfeng LinLLM ReliabilityLLM Evaluation

  7. MuSeToM: Multi-Sensory Evaluation of Belief Revision with Textual Cues Across Six Senses

    Nov 12, 2026Chunkit Chan, Kawai Chung, Sze Heng Douglas Kwok +13Theory of MindBelief Updating in LLMs

  8. Same Answer, Different Future: Dissociating Belief Content and Evidence Weight in Language Models

    Nov 5, 2026Weihan Li, Junhao Wu, Xinlei Chen +1Belief Updating in LLMsProbability Elicitation from Language Models

  9. Knowingly Wrong: Training LLMs Against Their Own Beliefs can Induce Misalignment

    Nov 3, 2026Ariana M. Villegas-Suarez, Guido Freire, Sohan Venkatesh +1Emergent Misalignment in Language ModelsKnowledge Conflicts in Language Models

  10. Do Language Models Know What Compelling Scientific Evidence Looks Like?

    Oct 29, 2026Conor Igoe, Gabriel Grand, Zane Koch +6Scientific Reasoning in Language ModelsEvidence-Grounded Reasoning

  11. Auditable Belief Updates as a Missing Safety Layer for LLM-Powered Multi-Agent Systems

    Oct 28, 2026Zhuoyun Li, Boxuan Wang, Jinwei Hu +2AI Agent SafetyLLM Agent Safety

  12. Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering

    Oct 26, 2026Duc-Vinh TranConversational QABelief Updating in LLMs

  13. The Patient Is Always Right? Evaluating Sycophancy in Medical Diagnosis with Graded Patient Beliefs

    Oct 25, 2026Viswonathan Manoranjan, Andrea Mock, Myra Cheng +4LLM SycophancyClinical Language Model Evaluation

  14. Do "True" and "False" Add Up to One? Complement Consistency in LLM-Agent Opinion Simulations

    Oct 23, 2026Anouar Imel, Jin QiBehavioral Consistency of LLM AgentsLarge Language Model-Based Social Simulation

  15. How Much Do You Trust A Dog? Quantifying Hint-Following in Reasoning Language Models

    Oct 23, 2026Anxin Guo, Chirag Pabbaraju, Sara Ahmadian +2LLM SycophancyLLM Reliability

  16. Invisibilization as a Source of Probability Bias in Language Models

    Oct 22, 2026Christian Arnold, Ernesto Calvo, Guadalupe Gonzalez +2Demographic Bias in Language ModelsLanguage Model Bias Evaluation

  17. Large language models show systematic failures in incorporating evolving medical knowledge

    Oct 22, 2026Min Zeng, Zaifu Zhan, Kai Yu +1Medical LLMsLLM Reliability

  18. From Prompts to Posteriors: Can LLM Agents Perform Bayesian Inference in Context?

    Oct 21, 2026Linus Filbry, Luckeciano C. Melo, Yarin GalBelief Updating in LLMsBayesian Inference

  19. To Manipulate or Not: Opinion Dynamics in LLM Interactions

    Oct 20, 2026Mohammad Siahkamari, Phuoc Dat Nguyen, Stephen MacNeilPersuasion in Language ModelsBelief Updating in LLMs

  20. Is Sycophancy a Form of In-Context Updating?

    Oct 20, 2026Itai Shapira, Myra Cheng, Hadas OrgadLLM SycophancyBelief Updating in LLMs

  21. LAMB: Latent Multi-Hypothesis Beliefs for Long-Horizon Large Language Agents

    Oct 18, 2026Peien Ding, Hui Li, Yuhan Liu +1Long-Horizon LLM AgentsMemory-Augmented Language Model Agents

  22. The One-Pixel of Reasoning: One Truthful Relation Is Enough to Mislead LLM Agents

    Oct 17, 2026Jinwei Hu, Youcheng Sun, Yi Dong +1Adversarial Attacks on LLMsLLM Agent Reliability

  23. Process Reward Modeling with Reliability- and Conservatism-Aware Bayesian Belief

    Oct 16, 2026Zhongjian Qiao, Zhixiang Zhou, Jiafei Lyu +4MLLM ReasoningProcess Reward Models

  24. Believed but Not Said: Measuring Hidden Belief Collapse in Multi-Agent LLM Discussions

    Oct 15, 2026Liangjie Huang, Henry Peng Zou, Wei-Chieh Huang +3Multi-Agent LLM SystemsLLM Agent Reliability