[KongchangAI]
Model

cross-encoder/ms-marco-MiniLM-L-6-v2

cross-encoder/ms-marco-MiniLM-L-6-v2 is a MiniLM-based cross-encoder model fine-tuned on the MS MARCO dataset for reranking retrieval results in RAG pipelines.

Timeline (last 90 days)

Oct 14

The local Cross-Encoder in the project uses the cross-encoder/ms-marco-MiniLM-L-6-v2 model, running locally based on Hugging Face Transformers and PyTorch.

Unverified50%
Oct 14

In the empirical tests, the MiniLM Cross-Encoder, biased toward lexical matching and sensitive to the scrambled text sequences produced by PDF parsing, pushed both correct answer chunks out of the Top 5, causing ranking degradation.

Unverified50%
Oct 12

cross-encoder/ms-marco-MiniLM-L-6-v2是一个常用的轻量级交叉编码器选择

Unverified50%

All Facts (3)

Source Articles