Unverified50% confidenceFactExact time
Ablation 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.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
Unverified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77% similarVerifiedInterpretability 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.76% similarUnverifiedVariance decomposition analysis showed that SCRIBE still explained unique brain-predictive information beyond the syntactic features selected by the researchers.75% similarUnverified多项实验表明，模型的表面推理与其实际决策机制之间存在显著偏差，通过激活修补可发现真正影响输出的内部路径与书面推理不吻合75% similarUnverifiedIn traditional mathematics, proving a long-standing conjecture stimulates a large amount of subsequent activity, forming an intergenerational relay mechanism of understanding72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1046855API
curl https://kongchang.com/api/v1/knowledge/claims/1046855MCP
get_claim(id=1046855)