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

Question Answering on WebQ (EM, ROUGE, F1)

32EM

GRIP w/o RL

Updated 4d ago

Evaluation Results

Method	Links
GRIP w/o RL 2026.04		32	39.9	35.1
GRIP 2026.04		31.4	39.3	34.6
DRAGIN 2026.04		25.2	31.5	35.7
FLARE 2026.04		24.2	30.3	34.7
GPT-4o 2026.04		23.5	32.3	37
R1-Searcher 2026.04		21.8	26.1	30.6
RetRobust 2026.04		21.8	27.4	31.7
RobustRAG 2026.04		21.5	25	29.1
Robust-RAG 2026.04		21.1	25.8	30
Robust-RAG 2026.04		19.8	24.5	28.7
InstructRAG 2026.04		19.5	26.8	31.4
R1-Searcher 2026.04		19.3	26.7	30.7
GRIP 2026.04		19.3	28.3	33.2
ETC 2026.04		18.9	26.6	30.4
SFT-RAG 2026.04		18.9	18.3	23.1
INFO-RAG 2026.04		18.1	27.7	28.9
GRIP 2026.04		16.8	27.5	32.3
GainRAG 2026.04		16.5	24.5	28.9
Self-RAG 2026.04		16.4	26.5	27.4
GPT-3.5 Turbo 2026.04		15.9	27.4	31.9
Single-RAG 2026.04		15.3	22.6	26.3
Instruct 2026.04		14.8	25.1	29.4
Instruct 2026.04		14.5	22.6	26.4
Single RAG 2026.04		14	22.7	26.6
R1-Searcher 2026.04		12.9	21.6	25.4
Single-RAG 2026.04		8.7	16.7	20.2
Instruct 2026.04		7.7	16.3	20