Unverified50% confidenceFactExact time
当模型对某个训练样本的预测概率接近1时，该样本对应的交叉熵损失趋近于零，模型几乎无法获得梯度更新信号
1
Sources
50%
Confidence
Long-term
Relevance
8/25/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe cross-entropy loss in classification tasks is essentially the negative of the log-likelihood function, and minimizing cross-entropy is equivalent to maximizing the log-likelihood of the model on the training data.77% similarUnverifiedAI生成假设的边际成本趋近于零，一旦模型训练完成每增加一条假设几乎不产生额外成本74% similarUnverifiedOpenAI's chart shows that at lower test-time compute levels, the model's success rate on this problem is close to zero, with occasional success only under extremely long reasoning chains.70% similarUnverified困惑度是测试集上每个token平均负对数似然的指数，值越低代表模型对语料的预测越准确70% similarUnverifiedFinite-sample estimation of entropy rate faces a systematic negative bias problem, where plug-in estimators tend to underestimate the true entropy; this can be compensated using the Miller-Madow correction by adding a (k-1)/(2N) term.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/797140API
curl https://kongchang.com/api/v1/knowledge/claims/797140MCP
get_claim(id=797140)