Unverified50% confidenceFactExact time
RRF uses only ranks rather than raw scores, avoiding the problem that the dimensions of dense vectors and sparse BM25 are not comparable.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEmbeddingGemma 2 inference must use BF16 or FP32 and should not use FP16, because activation values exceed the dynamic range of FP16, which can cause NaN values or quality degradation without any error being raised.69% similarUnverifiedUsing a 4B model to rerank the candidate shortlist filtered by BM25 outperformed BM25 alone (+0.047, confidence interval [0.020, 0.073]), even though the 4B reranker itself did not pass qualification.67% 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64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011858API
curl https://kongchang.com/api/v1/knowledge/claims/1011858MCP
get_claim(id=1011858)