Unverified50% confidenceFactExact time
The multiple testing problem refers to the phenomenon whereby the overall false positive rate rises sharply when multiple hypotheses are tested simultaneously; testing 20 capability dimensions at once means that on average 1 dimension will appear to show significant regression even if the agent has not changed.
1
Sources
50%
Confidence
Long-term
Relevance
9/27/2026
First Seen
Sources
Related Entities
Related Claims
Verified若用含测试集的全量数据计算均值和标准差，会导致预处理泄漏，使评估指标虚高，无法真实反映模型泛化能力76% similarVerifiedThe double descent phenomenon indicates that when the number of model parameters far exceeds the amount of data, the test error may actually decrease again.73% similarUnverified多Agent拆分能提升单点准确率，但引入交接判断准确性、上下文丢失风险和调试复杂度等新挑战73% similarVerifiedSome models carry the risk of over-optimization toward test sets, causing leaderboard scores to deviate from real-world performance (benchmark saturation).73% similarUnverifiedAgents face the problem of error cascading: judgment errors in early steps are amplified exponentially with the number of iterations.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/953024API
curl https://kongchang.com/api/v1/knowledge/claims/953024MCP
get_claim(id=953024)