List your company

Open-Domain Question Answering on TriviaQA (SubEM, F1)

74.01SubEMStable-RAG
44.640452.265259.8967.5148Jan 6, 2026

Evaluation results

MethodLinks
Stable-RAGBackbone=LLaMA3-8B-Instruct, Retriever=DPR, exhaustive decoding=true2026.0174.0174.12
Stable-RAGBackbone=LLaMA3-8B-Instruct, Retriever=DPR, exhaustive decoding=false2026.0173.4373.76
Stable-RAGBackbone=LLaMA3-8B-Instruct, Retriever=Contriever, exhaustive decoding=true2026.0172.1371.89
Stable-RAGBackbone=LLaMA3-8B-Instruct, Retriever=Contriever, exhaustive decoding=false2026.0172.0571.56
Stable-RAGBackbone=Qwen3-8B, Retriever=DPR, exhaustive decoding=true2026.0171.7473.4
Stable-RAGBackbone=Qwen3-8B, Retriever=DPR, exhaustive decoding=false2026.0171.3272.89
ATMBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0170.1270.35
ATMBackbone=Qwen3-8B, Retriever=DPR2026.0170.0671.67
Pos2DistillBackbone=Qwen3-8B, Retriever=DPR2026.0169.8771.35
Vanilla RAGBackbone=Qwen3-8B, Retriever=DPR2026.0169.6271.03
RetRobustBackbone=Qwen3-8B, Retriever=DPR2026.0169.1270.33
Ms-PoEBackbone=Qwen3-8B, Retriever=DPR2026.0169.0370.84
RetRobustBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0168.6770.42
RAATBackbone=Qwen3-8B, Retriever=DPR2026.0168.5469.88
RAATBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0168.0369.12
Vanilla RAGBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0167.1268.61
Stable-RAGBackbone=Qwen3-8B, Retriever=Contriever, exhaustive decoding=true2026.0167.1168.79
Stable-RAGBackbone=Qwen3-8B, Retriever=Contriever, exhaustive decoding=false2026.0166.5868.13
Pos2DistillBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0166.5768.12
Ms-PoEBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0166.4867.73
ATMBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0166.3767.12
RAATBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0165.5866.94
Pos2DistillBackbone=Qwen3-8B, Retriever=Contriever2026.0164.9566.81
Ms-PoEBackbone=Qwen3-8B, Retriever=Contriever2026.0164.8866.72
RetRobustBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0164.8566.72
ATMBackbone=Qwen3-8B, Retriever=Contriever2026.0164.7866.57
Vanilla RAGBackbone=Qwen3-8B, Retriever=Contriever2026.0164.3566.29
Ms-PoEBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0164.2166.14
Pos2DistillBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0164.1365.78
Vanilla RAGBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0163.8965.43
RetRobustBackbone=Qwen3-8B, Retriever=Contriever2026.0163.4965.39
RAATBackbone=Qwen3-8B, Retriever=Contriever2026.0163.1265.17
Vanilla SFTBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0157.152.51
Direct GenerationBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0155.9258.95
Direct GenerationBackbone=LLaMA3-8B-Instruct, Retriever=DPR2026.0155.9258.95
Vanilla SFTBackbone=LLaMA3-8B-Instruct, Retriever=Contriever2026.0155.5251.4
Vanilla SFTBackbone=Qwen3-8B, Retriever=DPR2026.0154.4650.17
Vanilla SFTBackbone=Qwen3-8B, Retriever=Contriever2026.0151.8747.62
Naive GenerationBackbone=Qwen3-8B, Retriever=Contriever2026.0145.7748.16
Naive GenerationBackbone=Qwen3-8B, Retriever=DPR2026.0145.7748.16