Unverified50% confidenceTradeoffExact time
基准测试面临饱和（Saturation）和数据污染（Data Contamination）两个系统性挑战，当多个主流模型得分超过95%时测试集失去区分能力
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
The Truth About "GPT-5.6": Debunking Fake Rumors and Third-Party Top-Up Scams
bilibili小小阿白6667/9/2026
Related Claims
Partially VerifiedBenchmark contamination refers to a model having already seen test questions during training, resulting in inflated scores that fail to reflect true generalization ability.74% similarUnverified多重比较问题指同时测试多个模型变体时，即使每个单独检验显著性水平设为0.05，整体假阳性率也会急剧膨胀72% similarUnverified对同一测试用例执行多次调用统计成功比例得到通过率，比二元的通过/失败判断更能反映模型真实表现69% similarVerifiedIn scenarios with extreme class imbalance, even an accuracy as high as 95% may indicate that the model is completely unable to identify the minority class.69% similarUnverifiedInfoNCE 损失通过在一批候选中最大化正样本对相似度、最小化负样本对相似度进行训练，温度超参数τ越小分布越集中68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/472661API
curl https://kongchang.com/api/v1/knowledge/claims/472661MCP
get_claim(id=472661)