Unverified50% confidenceFactExact time
Dai等人在2022年发表了论文《Knowledge Neurons in Pretrained Transformers》，通过定位和编辑特定神经元可以修改模型对特定事实的记忆
1
Sources
50%
Confidence
Long-term
Relevance
8/28/2026
First Seen
Sources
Related Entities
Related Claims
Unverified线性注意力思路由 Katharopoulos 等人在 2020 年论文《Transformers are RNNs》中系统化70% similarUnverifiedTransformer原始论文于2017年由Google Brain团队发表67% similarUnverifiedAnthropic published a paper on superposition in 2022, revealing the phenomenon of individual neurons simultaneously encoding multiple concepts.62% similarUnverifiedThe field of mechanistic interpretability was largely founded by the Anthropic research team (especially Chris Olah) around 2021 through the "Transformer Circuits" series of papers.62% similarUnverifiedAnthropic 发布了《Transformer Circuits》系列文章59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/810526API
curl https://kongchang.com/api/v1/knowledge/claims/810526MCP
get_claim(id=810526)