The belief-action gap in language model agents
cs.LG, cs.AI, cs.CL
Submitted: 2026-10-12
Updated: 2026-10-12
Comments: 25 pages, 10 figures; includes appendices
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- gpt-oss-120b & gpt-oss-20b Model Card
- Understanding intermediate layers using linear classifier probes
- Concrete Problems in AI Safety
- Eliciting Latent Predictions from Transformers with the Tuned Lens
- Thought Anchors: Which LLM Reasoning Steps Matter?
- Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought
- Thinking Past the Answer: Evaluating Harmful Overthinking in Large Reasoning Models
- Answer-Distribution Trajectories: A Stochastic-Dynamics View of LLM Reasoning
- Emergent Response Planning in LLMs
- Understanding and Mitigating Premature Confidence for Better LLM Reasoning
- Early Stopping for Large Reasoning Models via Confidence Dynamics
- Measuring Faithfulness in Chain-of-Thought Reasoning
- Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
- Aligning Large Language Model Agents with Rational and Moral Preferences: A Supervised Fine-Tuning Approach
- Beyond the Commitment Boundary: Probing Epiphenomenal Chain-of-Thought in Large Reasoning Models
- Steering Language Models With Activation Engineering
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks