Unverified50% confidenceBenchmarkExact time
The trained model achieves a per-token negative log-likelihood (NLL) of 0.93, compared to 12.60 for a randomly initialized twin model of the same architecture.
1
Sources
50%
Confidence
Long-term
Relevance
9/23/2026
First Seen
Sources
Related Entities
Related Claims
Unverified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70% similarUnverifiedFor negative control models trained only on legitimate training sets, the measurement method yields an effect value of -0.012 and has never falsely reported contamination.67% similarUnverifiedFine-tuned DistilBERT achieves F1=0.9993 on in-distribution testing, but loses 12.9 mean average precision points under topic shift, a greater degradation than the linear model's 5.2 points.67% similarUnverifiedMcNemar检验的零假设是b和c在总体中相等，即两个模型犯错方向对称，无系统性优劣67% similarUnverifiedIn models trained in practice, the weight distribution is often non-uniform, with a large number of weights tending toward zero, forming natural sparsity.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/942824API
curl https://kongchang.com/api/v1/knowledge/claims/942824MCP
get_claim(id=942824)