Unverified50% confidenceFactExact time
When generating new tokens using a KV cache, it is necessary to compute the dot product between the query vector and all historical key vectors in the cache, with a complexity of O(N).
1
Sources
50%
Confidence
Long-term
Relevance
9/6/2026
First Seen
Sources
Related Entities
Related Claims
Unverified标准多头注意力中每个Token需与序列中所有其他Token计算相关性权重，计算复杂度为O(N²)70% similarUnverifiedThe computational complexity of cross-encoders is O(n²); document representations cannot be precomputed and scoring must be done online pair by pair, so cross-encoders are typically only used to rerank the top-k results returned by an initial retrieval stage.65% similarUnverified非原生图数据库需通过全局索引查找相邻节点，时间复杂度随图规模增长为O(log n)甚至O(n)62% similarUnverified自注意力通过将输入序列映射为查询Q、键K、值V三组向量，计算Q与K的点积相似度后经Softmax归一化得到注意力权重，再加权汇聚V向量60% similarUnverifiedWHERE id > maxId LIMIT N的游标分页方式（Keyset Pagination）利用主键索引的B+树有序性，时间复杂度始终为O(log N + M)58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/865955API
curl https://kongchang.com/api/v1/knowledge/claims/865955MCP
get_claim(id=865955)