Confident at the moment of action: belief miscalibration in LLM play under hidden information
cs.AI, cs.CL, cs.LG
Submitted: 2026-08-25
Updated: 2026-08-25
Terminology
Sources
- Language Models (Mostly) Know What They Know
- Same Answer, Different Confidence: Protocol Sensitivity in LLM Confidence Calibration
- ConfTuner: Training Large Language Models to Express Their Confidence Verbally
- Calibrating Verbalized Confidence with Self-Generated Distractors
- Taming Overconfidence in LLMs: Reward Calibration in RLHF
- ConfidenceBench: Evaluating Confidence Calibration in Large Language Models
- Agent-BRACE: Decoupling Beliefs from Actions in Long-Horizon Tasks via Verbalized State Uncertainty
- BayesBench: Evaluating LLM Belief Trajectories Under Multi-Turn Evidence Accumulation
- Knowing What You Know Is Not Enough: Large Language Model Confidences Don't Align With Their Actions
- WOLF: Werewolf-based Observations for LLM Deception and Falsehoods
- MafiaScope: Non-Invasive, Time-Resolved Belief Probing for LLM Agents in Social Deduction Games
- Readable Minds: Emergent Theory-of-Mind-Like Behavior in LLM Poker Agents
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems
- VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection