List your company

Multi-hop Question Answering on HotpotQA (EM, F1, LLM)

61.13EMSeleCom
-1.457214.791431.0447.2886Jan 25, 2026

Evaluation results

MethodLinks
SeleComBackbone=Mistral-7B-Instruct, Top-k=5, Comp. Rate=82x2026.0161.1376.1676.57
SeleComBackbone=Qwen2.5-7B-Instruct, Top-k=5, Comp. Rate=82x2026.0159.1474.3875.43
LLM (with RAG*)Backbone=Mistral-7B-Instruct, Top-k=5, Comp. Rate=-2026.0157.2274.3669.96
LLM (with RAG*)Backbone=Qwen2.5-7B-Instruct, Top-k=5, Comp. Rate=-2026.0156.2672.7774.94
SeleComBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=82x2026.0144.4657.8857.49
LLM (with RAG*)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.0143.1952.6151.71
LLM (with RAG*)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.0142.9256.4355.98
SeleComBackbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=82x2026.0142.3555.3455.91
LLM (with RAG)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.0136.9649.9557.17
COCOMBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=128x2026.0125.935.3634
ICAEBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=4x2026.0124.7740.0653.82
LLM (with RAG)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.0118.6935.8758.57
RECOMPBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=N/A2026.0118.5135.7656.4
LLM (w/o RAG)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.0115.7323.7924.63
LLMLingua-2Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=4x2026.0112.4626.8645.91
xRAGBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=164x2026.018.4522.5245.65
LLM (w/o RAG)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.015.9716.2829.97
PISCOBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=64x2026.010.9513.0854.18