Unverified50% confidenceBenchmarkExact time
On a four-model hidden Markov family, learning an unknown Bayesian decision requires Θ(l⁻²) records.
1
Sources
50%
Confidence
Long-term
Relevance
10/4/2026
First Seen
Sources
Related Entities
Related Claims
Unverified许多开源模型通过对GPT-4、Claude等闭源模型的知识蒸馏训练而来，这可能导致其继承教师模型的自信偏差，从而不愿在工具调用错误时自我纠正64% similarUnverifiedDiscriminative models learn which class a given input-output belongs to or its probability, with typical representatives including BERT and its fine-tuned variants.64% similarUnverifiedA Markov decision process is composed of a five-tuple (S, A, P, R, γ), with the Markov property at its core64% similarUnverified研究者发现当模型参数量跨越某一阈值时，代码推理能力会出现非线性跃升，这一现象在GPT-4、Claude 3等模型上均有体现63% similarVerified混合专家架构MoE通过稀疏激活机制，推理时只激活模型总参数的1/8至1/463% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/971862API
curl https://kongchang.com/api/v1/knowledge/claims/971862MCP
get_claim(id=971862)