Unverified50% confidenceFactExact time
Vector retrieval finds the most relevant chunks by computing the cosine similarity or Euclidean distance between the query vector and document chunk vectors.
1
Sources
50%
Confidence
Long-term
Relevance
9/19/2026
First Seen
Sources
Related Entities
Related Claims
Unverified检测输入分布漂移的常用方法包括对输入文本生成 embedding 向量后计算统计距离（如 KL 散度、Wasserstein 距离）、监控主题分布变化、追踪输入长度复杂度特征以及监控裁判分数分布偏移68% similarUnverified质心追踪算法通过计算当前帧质心与上一帧已注册目标质心的欧氏距离进行配对，计算复杂度随目标数量呈线性甚至平方级增长65% similarUnverifiedThe reranking module typically uses a cross-encoder model to compute relevance scores between queries and documents, performing a secondary ranking of the candidate documents from the initial retrieval.64% similarUnverified对于相似度处于灰色地带（如0.85到0.95之间）的查询，可触发轻量级重排序或人工规则校验再决定是否命中，实现置信度分层63% similarUnverified梯度检验用数值方法(f(x+ε)-f(x-ε))/(2ε)近似计算梯度，与解析梯度对比以确保反向传播代码正确性63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/934422API
curl https://kongchang.com/api/v1/knowledge/claims/934422MCP
get_claim(id=934422)