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

Open-domain Question Answering on Natural Questions (F1, AR)

52.2F1 Score

UMGP-RAG

Updated 1mo ago

Evaluation Results

Method	Links
UMGP-RAG 2026.06		52.2	49.84
LongRAG 2026.06		51.45	49.49
UMGP-RAG 2026.06		51.28	48.78
REPLUG 2026.06		51.2	46.42
LongRAG 2026.06		51.13	49.1
MS-PoE 2026.06		50.46	47.2
MS-PoE 2026.06		50.05	47.09
Hybrid 2026.06		49.86	47.66
UMG-RAG 2026.06		49.74	47.9
REPLUG 2026.06		49.62	44.59
RAG 2026.06		49.41	45.93
RAG 2026.06		49.39	46.21
RAG 2026.06		49.23	47.41
Hybrid 2026.06		49.16	47.41
UMG-RAG 2026.06		48.73	47.16
UMGP-RAG 2026.06		48.11	45.33
LongLLMLingua 2026.06		46.31	44.62
LongLLMLingua 2026.06		45.61	43.46
UMG-RAG 2026.06		45.14	43.04
Hybrid 2026.06		45.04	42.93
PINE 2026.06		44.52	46.03
PINE 2026.06		42.91	44.38
Summarized RAG 2026.06		35.37	31.62
Summarized RAG 2026.06		35.2	31.27
LongLLMLingua 2026.06		25.73	24.29
MS-PoE 2026.06		24.35	21.92
RAG 2026.06		24.14	21.78
LongRAG 2026.06		23.18	20.94
REPLUG 2026.06		22.54	18.96
PINE 2026.06		20.84	20.97
Summarized RAG 2026.06		19.72	17.03