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

Text-based Question Answering on HotpotQA

21Exact Match (EM)

Full Context

Updated 1mo ago

Evaluation Results

Method	Links
Full Context 2026.06		21	44.9	-	1,701
Qwen3-Emb-0.6B-ft 2026.06		18.6	42.5	80.5	267
Latent Memory 2026.06		17.5	37.8	86.6	76
Qwen3-Emb-0.6B-ft 2026.06		16.3	38.4	59.7	127
BM25 Retrieval 2026.06		15.9	37.8	65.5	255
Qwen3-Emb-0.6B 2026.06		15.1	36.6	70.1	256
Latent Memory 2026.06		14.8	34.8	62.2	46
Dense Retrieval 2026.06		13.9	34	62.4	244
BM25 Retrieval 2026.06		13.8	33.8	45.5	120
Qwen3-Emb-0.6B 2026.06		12.6	32.7	50.1	124
Qwen3-Emb-0.6B-ft 2026.06		12.3	32.2	37.9	80
Latent Memory 2026.06		12.3	30.4	34.8	36
xRAG 2026.06		12.1	25.3	-	42
Dense Retrieval 2026.06		11.3	30.2	42.2	117
BM25 Retrieval 2026.06		11	29.6	30.2	76
Qwen3-Emb-0.6B 2026.06		10.3	28.4	33.2	79
Dense Retrieval 2026.06		9	26.4	27.9	75
LLMLingua 2026.06		8.9	27.6	-	406
LLMLingua 2026.06		6.8	22.6	-	217
CLaRa-16x 2026.06		5.2	15.3	55.8	119
LLMLingua 2026.06		4.7	17.9	-	117