Uncertainty calibration
Technique · Safety & Alignment · 1 item since Jul 2026
Training models so their stated confidence matches how often they are right.
Trends
Items
- Jul 3
RLMF trains on metacognitive feedback
Reinforcement learning on a model's own self-judgments aligns expressed with intrinsic uncertainty, reaching state-of-the-art faithful calibration and beating standard RL by up to 63% without losing accuracy.