Unverified50% confidenceFactExact time
The Φ value of standard feedforward neural networks approaches zero, because information flows only in one direction and lacks recurrent integration.
1
Sources
50%
Confidence
Long-term
Relevance
10/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn non-convex small neural networks, the directional misalignment between Adam and NGD reaches an order of magnitude of approximately 10^3.67% similarUnverifiedThe model's output tokens are zero, because it only returns probabilities rather than generating text.64% similarUnverifiedProbes trained on fewer than 1% of neurons retain the vast majority of the detection accuracy achieved with the full feature set.64% similarUnverifiedIf a difficult problem fails across all samples (pass@k ≈ 0), the reward variance is zero, the gradient signal nearly vanishes, and the parameter update approaches zero.60% similarUnverified当权重仅为-1、0、+1时，矩阵乘法可完全用加减法代替，理论上可将神经网络推理功耗降低5-10倍59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/979371API
curl https://kongchang.com/api/v1/knowledge/claims/979371MCP
get_claim(id=979371)