Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

RQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Question AnsweringRQA
ASR88.2
130
Retrieval Question AnsweringRQA
Accuracy76
72
Question AnsweringRQA (test)
Accuracy79
60
Question AnsweringRQA MC
RACC (Accuracy)77.2
58
Robust Question AnsweringRQA Evolving evidence streams GPT-4o (test)
Accuracy72.68
24
RAG Poisoning Attack MitigationRQA
ASR (PIA)1
15
Question AnsweringRQA poison @ Position 10, k=10 (test)
Robustness Accuracy76
15
Question AnsweringRQA (poison @ Position 1, k=10) (test)
Robustness Accuracy0.7
15
RAG RobustnessRQA
Paradox RACC66.4
12
RAG RobustnessRQA-MC
Paradox RACC80
12
Short-answer QARQA
Accuracy71
8
Short-form open-domain QARQA
PIA Racc Score72
6
Multiple-choice QARQA-MC
Accuracy81
6
Showing 13 of 13 rows