[KongchangAI]
Unverified50% confidenceOpinionExact time

蒸馏训练的教师模型输出极少包含自我纠错示例,导致学生模型形成「强模型不犯错」的隐性信念,在校验失败时选择重复而非修正

1
Sources
50%
Confidence
Long-term
Relevance
7/6/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/114525
API
curl https://kongchang.com/api/v1/knowledge/claims/114525
MCP
get_claim(id=114525)