Beyond Verbalized Confidence: Calibrating Reasoners with Differentiable Readouts
cs.LG, cs.AI, cs.CL
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/huggingface/Math-Verify
Terminology
Sources
- Uncalibrated Reasoning: GRPO Induces Overconfidence for Stochastic Outcomes
- Evaluating Large Language Models Trained on Code
- Language Models (Mostly) Know What They Know
- C$^2$GSPG: Confidence-calibrated Group Sequence Policy Gradient towards Self-aware Reasoning
- Reasoning about Uncertainty: Do Reasoning Models Know When They Don't Know?
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Qwen3 Technical Report
- On Verbalized Confidence Scores for LLMs
- LaSeR: Reinforcement Learning with Last-Token Self-Rewarding
- Know What You Know: Metacognitive Entropy Calibration for Verifiable RL Reasoning
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: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks