Verified65% confidenceFactExact time
The maximum value of the derivative of the Sigmoid function is only 0.25, leading to the vanishing gradient problem in deep networks.
3
Sources
65%
Confidence
Long-term
Relevance
7/23/2026
First Seen
Sources
Related Claims
UnverifiedThe Sigmoid function compresses the input to between 0 and 1, with a maximum slope of only 0.25.77% similarUnverifiedWhen the cumulative sum reaches 1.0, the smallest change representable by FP32 is 1 ULP = 2⁻²³ ≈ 0.00000012.57% similarUnverifiedThe same model's scores on DeepCWE range from 65.5 to 74.2 depending on the agent framework used; the official harness uses the upper end of the range, while under Claude Code's harness the score is only 69.8.56% similarUnverifiedOn 1024-step sequences, the average gradient of the initial token for PLUG and ILRM remains stably at around the 10⁻⁴ order of magnitude, while the standard GRU gradient drops as low as approximately 10⁻¹⁴⁵, resulting in numerical underflow.56% similarUnverifiedINT4量化将原本由65536个不同值（FP16）表示的权重压缩到仅16个离散级别56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/601149API
curl https://kongchang.com/api/v1/knowledge/claims/601149MCP
get_claim(id=601149)