Unverified50% confidenceFactExact time
Min-K% 方法通过检查模型在文本中最低概率token的平均对数概率来判断样本是否为训练集成员
1
Sources
50%
Confidence
Long-term
Relevance
8/16/2026
First Seen
Sources
Terminal Bench 3发布：未被污染的AI终端能力基准测试
Related Claims
UnverifiedResearch shows that on certain open-source benchmarks, the n-gram overlap rate between the training data of leading large models and the evaluation data can be as high as 15%–30%.66% similarUnverifiedword2vec负采样中，对每个正样本随机采样k个（通常5-20个）负样本，采样概率通常按词频的3/4次方分布66% similarUnverified负采样将每步训练的复杂度从O(V)降低到O(k)，k是负样本数（通常取5-20）65% similarUnverified在小样本LoRA训练中，rank值设置存在偏差-方差权衡，实践中rank=4至32是常见的安全区间65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/757585API
curl https://kongchang.com/api/v1/knowledge/claims/757585MCP
get_claim(id=757585)