Unverified50% confidenceBenchmarkExact time
Using 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.
1
Sources
50%
Confidence
Long-term
Relevance
10/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAfter testing 30 layer combinations, the best pairing was SenseNova L32 mapped to MiniMax H3 L49, using a 4096→128→5120 projector that achieved a cosine similarity of 0.9042 on the validation set.72% similarUnverified生产中典型做法是先用向量检索和BM25快速召回50-100个候选结果，再用reranker精排取top-k送入LLM69% similarUnverifiedXingchen 4.0 29B is a domestically developed open-source model that adopts a MoE architecture with 29B total parameters, activating only 4B parameters at a time.69% similarUnverifiedThe core mechanism of BM25 is based on three factors: term frequency (TF), inverse document frequency (IDF), and document length normalization, with parameter k1 typically set between 1.2 and 2.0 and b typically set at 0.75.68% similarUnverifiedThe author noted that other models previously tested, such as 35B and 3B variants, actually performed worse than the Q2 version of Qwen3 27B on the Kanban task.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/966724API
curl https://kongchang.com/api/v1/knowledge/claims/966724MCP
get_claim(id=966724)