Unverified50% confidenceFactExact time
2022年Meng等人的「知识神经元」研究表明，事实性知识主要存储在MLP层的键值对中，通过修改特定MLP层的参数可以精确编辑模型的知识
1
Sources
50%
Confidence
Long-term
Relevance
8/27/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the Transformer architecture, factual knowledge is considered to be primarily distributed in the weight matrices of MLP (feed-forward network) layers, while the attention mechanism is responsible for dynamically retrieving and integrating this knowledge during inference.74% similarUnverifiedLarge models store parametric knowledge, with world knowledge compressed and encoded into billions of neural network weights rather than being stored in a structured knowledge base that can be precisely retrieved.73% similarUnverifiedAnthropic's mechanistic interpretability research attempts to locate the storage positions of specific knowledge at the neural network activation level.71% similarUnverifiedMLP层充当模型的长期事实记忆库，其键值对是预训练阶段固化在权重中的静态参数69% similarUnverifiedThe core data model of a knowledge graph is the triple (subject–predicate–object).68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/809211API
curl https://kongchang.com/api/v1/knowledge/claims/809211MCP
get_claim(id=809211)