Verified85% confidenceFactExact time
A perfectly calibrated model achieves an accuracy of exactly 70% across all predictions for which it claims 70% confidence.
5
Sources
85%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
元认知反馈：用强化学习让大模型准确表达不确定性
hackernewshackernews7/7/2026
Related Claims
UnverifiedWhen a well-calibrated person says 'I am 70% confident,' the actual success rate of such judgments should be close to 70%.78% similarVerifiedIn a well-calibrated model, an output score of 0.8 should statistically correspond to an actual accuracy of 80%.77% similarVerifiedA "well-calibrated" probability judgment means that when a model says something has a 70% chance of occurring, the actual occurrence rate of that event across many similar predictions is close to 70%.76% similarUnverifiedOn reasoning-based question-answering tasks, the strongest closed-source models achieved accuracy exceeding 90%.73% similarVerifiedModel calibration is defined in statistics as: if a model claims a certain answer has 70% confidence, then the actual proportion of correct answers among a large number of such claims should be close to 70%.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/261060API
curl https://kongchang.com/api/v1/knowledge/claims/261060MCP
get_claim(id=261060)