Blended RAG: Improving RAG (Retriever-Augmented Generation) Accuracy with Semantic Search and Hybrid Query-Based Retrievers
Open paper
Benchmark
Model
Rank
Results
open-domain-question-answering-on-squad1-1-dev
Blended RAG
#2
EM: 57.63
question-answering-on-natural-questions
Blended RAG
#14
EM: 42.63
question-answering-on-nq-beir
Blended RAG
#1
nDCG@10: 0.67
Rank counts only results with a code link.