RocketQA

Reported on 2 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing2 results

Information RetrievalonNatural Questions
Precision@100· 2020-10-16
88.5
best: 90.4 (ReAtt)
SOTA
RocketQA: An Optimized Training Approach to Dense Passage Retrieval for Open-Domain Question Answering arXiv:2010.08191
Information RetrievalonNatural Questions
Precision@20· 2020-10-16
82.7
best: 86 (ReAtt)
SOTA
RocketQA: An Optimized Training Approach to Dense Passage Retrieval for Open-Domain Question Answering arXiv:2010.08191