Unverified50% confidenceFactExact time
The GK summary (Greenwald-Khanna) uses counters with error bounds to provide deterministic approximate guarantees for quantiles.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedNumerical gradient checking approximates gradients using the central difference formula (f(θ+ε) - f(θ-ε)) / (2ε) and compares them with analytical gradients; a relative error below 1e-5 is considered correct, while an error above 1e-3 almost certainly indicates a bug.65% similarUnverifiedThe value of K in Top-K should be determined through offline evaluation by measuring Recall@K at different K values on an annotated test set and finding the inflection point where the recall curve begins to flatten.63% similarUnverifiedK折交叉验证的K通常取5或10，这在偏差-方差权衡上被实证为较优选择63% similarUnverifiedK-Quant是llama.cpp生态中的混合精度量化方案，根据每层权重敏感度分配不同量化bit数，可在相同平均bit数下将精度损失降低40-60%62% similarUnverifiedPrecision@k is defined as: the number of expected text chunks that were retrieved divided by k.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1043400API
curl https://kongchang.com/api/v1/knowledge/claims/1043400MCP
get_claim(id=1043400)