Unverified50% confidenceFactExact time
Cherry-picking is a pervasive risk in AI model evaluation and systematically overestimates model capabilities.
1
Sources
50%
Confidence
Long-term
Relevance
10/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAI model over-refusal is a known issue in safety alignment research, where RLHF and Constitutional AI training cause models to incorrectly classify harmless requests as requiring refusal75% similarUnverified在AI安全领域，失准指模型的实际行为偏离设计者或使用者意图的现象，当代大模型的失准更多表现为目标过度追求75% similarUnverifiedAI图像生成领域存在精选偏差（Cherry-Picking Bias），用户倾向于只展示最成功的生成结果，导致外界对模型能力产生过于乐观的估计74% similarUnverifiedStatic cognition of AI capabilities leads to two typical misjudgments: underestimation and overestimation.74% similarUnverified主动要求AI引入可信瑕疵可以对抗模型的讨好性偏差，从而提升输出的真实度73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/968601API
curl https://kongchang.com/api/v1/knowledge/claims/968601MCP
get_claim(id=968601)