Unverified50% confidenceFactExact time
After retraining Inception V1 with its original architecture and adding modern regularization, about 70% of the neurons in the last layer can be interpreted.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Neural Networks Were Always Clean? Goodfire Researcher Overturns Superposition Theory
bilibili迦太基盐商10/16/2026
Related Entities
Related Claims
UnverifiedXavier initialization adjusts the variance of weights according to the number of input and output neurons, and is suitable for Sigmoid/Tanh-type activation functions.63% similarVerifiedEach Transformer layer contains two core sub-modules: multi-head self-attention and a feed-forward neural network.61% similarUnverifiedThe most basic neural network consists of three parts: an input layer, a hidden layer, and an output layer.61% similarUnverifiedCamerata proposed the concept of "recursive generalization": neurons branch and grow along related directions during training, and each branch has a provenance rather than being forcibly stuffed in as superposition.60% similarUnverifiedNeural networks exhibit permutation symmetry—reordering the neurons in a given layer leaves the network's functionality completely unchanged, as long as the weights of adjacent layers are adjusted accordingly.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1052466API
curl https://kongchang.com/api/v1/knowledge/claims/1052466MCP
get_claim(id=1052466)