Unverified50% confidenceDecision RuleExact time
In the experimental results section, AI tends to list data in a flat manner, whereas the correct approach is to emphasize only the results that prove the core contribution points.
1
Sources
50%
Confidence
Long-term
Relevance
9/27/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCurrent mainstream AI scientist evaluation benchmarks almost exclusively focus on final outputs while discarding the reasoning process.76% 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75% similarUnverifiedAI在缺乏明确约束时倾向于用统计上最合理的内容填充空白，而非主动询问澄清，这是幻觉问题的延伸75% similarUnverifiedEvaluating the mathematical capabilities of AI should focus more on performance on out-of-distribution problems rather than on high benchmark scores75% 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75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/953045API
curl https://kongchang.com/api/v1/knowledge/claims/953045MCP
get_claim(id=953045)