Unverified50% confidenceFactExact time
19.4% (6,251 articles) of duplicate document contamination exists in the naive test set split.
1
Sources
50%
Confidence
Long-term
Relevance
9/23/2026
First Seen
Sources
Related Entities
Related Claims
Unverified研究表明即使是轻微的数据污染也可能导致评测分数虚高5-15个百分点67% similarUnverified实验采用8:2比例划分数据，20%作为内部测试集，训练集内部使用5折交叉验证，并移除重复记录以避免数据泄漏66% similarUnverifiedIn experiments on 622 public cases, reports selected by ranking on visible lexical scores achieved a composite score of 0.2909.65% similarUnverifiedOut of a total of 188 model × benchmark combinations, only 3 cases exhibit differential contamination cross-validated by two reference points.65% similarUnverifiedIn the experiment, after prompts exceeded 50, the human interception rate dropped sharply from 17% to 5%, reflecting the phenomenon of approval fatigue.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/942803API
curl https://kongchang.com/api/v1/knowledge/claims/942803MCP
get_claim(id=942803)