Unverified50% confidenceFactExact time
Because expert pruning preserves the attention architecture and KV cache format, the KV cache handoff requires no cache conversion, and the runtime only needs to switch weight masks between the two stages.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedKV Cache only caches K and V but not Q, because Q is only used in the current step and discarded after use, whereas historical K and V are re-accessed at every subsequent step.71% similarVerifiedKV Cache avoids redundant computation by caching already-computed Key and Value matrices, and advanced techniques such as PagedAttention further optimize cache memory management efficiency.71% similarUnverified上下文越长，模型需要缓存的 KV Cache 越大，KV Cache 使模型生成新 Token 时无需重新计算所有历史 Token 的注意力权重71% similarUnverifiedFP8 KV cache is a quantization technique that compresses the Key-Value cache in the attention mechanism from FP16 or BF16 to 8 bits, which can halve the GPU memory usage of the KV cache70% similarUnverifiedPrompt缓存（KV Cache）在相同系统提示多次使用时可复用缓存的KV矩阵，通常将延迟降低40-60%、成本降低50-90%70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1042670API
curl https://kongchang.com/api/v1/knowledge/claims/1042670MCP
get_claim(id=1042670)