Unverified90% confidenceFactTime unknown
PR曲线在样本不平衡场景下更有价值，根本原因在于精确率的计算不涉及真负例（TN）的数量
1
Sources
90%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
Unverified优质Prompt在大多数场景下比微调更高效，在数据量不足或需求模糊的阶段贸然做微调往往事倍功半70% similarUnverified西尔维斯特-加莱定理依赖于实数域的特殊性质，并非纯粹的组合结论，在复平面上失效68% similarUnverified漏洞数量激增并不必然意味着模型更不安全，可能是审查强度提升带来的统计偏差65% similarUnverifiedWilson置信区间相比简单正态近似在小样本场景下更为准确，能有效避免通过率接近0或1时出现区间越界的问题65% similarUnverifiedIn class-imbalanced scenarios, accuracy alone can be misleading and needs to be combined with a confusion matrix for more fine-grained evaluation.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/20877API
curl https://kongchang.com/api/v1/knowledge/claims/20877MCP
get_claim(id=20877)