[KongchangAI]
Unverified50% confidenceFactExact time

Abliteration技术源自2024年开源社区对大语言模型内部表征的研究,理论基础是线性表征假说

1
Sources
50%
Confidence
Long-term
Relevance
7/20/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/568112
API
curl https://kongchang.com/api/v1/knowledge/claims/568112
MCP
get_claim(id=568112)