Unverified50% confidenceFactExact time
Without bias, when all inputs are zero the neuron output is also zero, which limits the model's expressive power.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedZero让人类只需提出期望的结果,然后在需要时审阅可读的代码投影72% similarUnverifiedAll-zero initialization causes a symmetry problem: all neurons in the same layer compute identical gradients and make identical updates, making it impossible to break symmetry.69% similarUnverifiedGreedy decoding at temperature zero further suppresses the probability of a model outputting ESCALATE, because models tend to select the answer pattern that appeared most frequently during training, and escalation is inherently rare in instruction fine-tuning data.68% similarUnverifiedThe TopK sparse autoencoder retains only the K neurons with the highest activation values during each forward pass, setting all others to zero.67% similarUnverifiedAfter removing the gating condition from the metacognitive prompt pipeline, accuracy fell to zero, reverting to the 5/17 baseline.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011887API
curl https://kongchang.com/api/v1/knowledge/claims/1011887MCP
get_claim(id=1011887)