Unverified50% confidenceFactExact time
Polysemanticity means that a single neuron or direction often encodes multiple unrelated concepts at once, creating a fundamental difficulty for reading out the model's internal meaning.
1
Sources
50%
Confidence
Long-term
Relevance
11/6/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedInterpretability research from MIT and Anthropic found that there are large numbers of polysemantic features inside models, with a single neuron potentially responding simultaneously to multiple seemingly unrelated semantic concepts.77% similarUnverified分词特征差异发生在推理链路底层，难以通过提示工程掩盖，可作为模型识别手段73% similarUnverified模糊的感受类描述会让大语言模型在多个可能的方向上随机游走，而具体的问题描述则直接指向解决路径70% similarUnverifiedAblation experiments on semantic attributes revealed that semantic contributions span multiple dimensions, producing cumulative effects on the predicted variance of subjects' brains rather than being dominated by a single dimension.70% similarUnverifiedThe core idea of the superposition hypothesis is: when the number of concepts a model needs to represent exceeds the dimensionality of its representation space, neural networks tend to encode these concepts using non-orthogonal dense directions.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1142304API
curl https://kongchang.com/api/v1/knowledge/claims/1142304MCP
get_claim(id=1142304)