List your company

Long-form Question Answering on Biography MixP (50% polluted)

70.11VeriScore F1@kMIRAGE
13.076427.883242.6957.4968Jul 6, 2026

Evaluation results

MethodLinks
MIRAGEBackbone=GPT-4o-mini2026.0770.11
MIRAGEBackbone=Qwen3-8B2026.0761.25
No-RAGBackbone=GPT-4o-mini2026.0758.99
MADAMRAGBackbone=GPT-4o-mini2026.0758.39
AstuteRAGBackbone=GPT-4o-mini2026.0756.55
BRIDGEBackbone=GPT-4o-mini2026.0756.37
SelfRAGBackbone=GPT-4o-mini2026.0756.23
No-RAGBackbone=Qwen3-8B2026.0753.75
DRAGINBackbone=GPT-4o-mini2026.0752.58
FLAREBackbone=GPT-4o-mini2026.0751.58
RetRobust-NLIBackbone=GPT-4o-mini2026.0749.5
MADAMRAGBackbone=Qwen3-8B2026.0748.91
RARGBackbone=GPT-4o-mini2026.0748.76
RobustRAGBackbone=Qwen3-8B2026.0748.14
TrustRAGBackbone=GPT-4o-mini2026.0748.03
RAGBackbone=GPT-4o-mini2026.0742.56
RobustRAGBackbone=GPT-4o-mini2026.0740.47
FLAREBackbone=Qwen3-8B2026.0736.77
SelfRAGBackbone=Qwen3-8B2026.0735.18
RARGBackbone=Qwen3-8B2026.0734.03
TrustRAGBackbone=Qwen3-8B2026.0733.64
DRAGINBackbone=Qwen3-8B2026.0733.36
BRIDGEBackbone=Qwen3-8B2026.0731.44
AstuteRAGBackbone=Qwen3-8B2026.0728.86
RAGBackbone=Qwen3-8B2026.0726.64
RetRobust-NLIBackbone=Qwen3-8B2026.0715.27