List your company

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

90.33VeriScore F1@kMIRAGE
55.011664.180873.3582.5192Jul 6, 2026

Evaluation results

MethodLinks
MIRAGEBackbone=GPT-4o-mini2026.0790.33
MIRAGEBackbone=Qwen3-8B2026.0787.56
No-RAGBackbone=GPT-4o-mini2026.0786.9
No-RAGBackbone=Qwen3-8B2026.0784.61
AstuteRAGBackbone=GPT-4o-mini2026.0781.14
BRIDGEBackbone=GPT-4o-mini2026.0780.71
DRAGINBackbone=GPT-4o-mini2026.0773.41
FLAREBackbone=GPT-4o-mini2026.0773.32
SelfRAGBackbone=GPT-4o-mini2026.0773.08
DRAGINBackbone=Qwen3-8B2026.0771.48
BRIDGEBackbone=Qwen3-8B2026.0770.5
RetRobust-NLIBackbone=GPT-4o-mini2026.0770.27
RobustRAGBackbone=Qwen3-8B2026.0769.21
AstuteRAGBackbone=Qwen3-8B2026.0767.22
RobustRAGBackbone=GPT-4o-mini2026.0766.22
TrustRAGBackbone=GPT-4o-mini2026.0765.66
MADAMRAGBackbone=GPT-4o-mini2026.0765.48
RARGBackbone=GPT-4o-mini2026.0765.38
SelfRAGBackbone=Qwen3-8B2026.0765.32
RetRobust-NLIBackbone=Qwen3-8B2026.0765.15
MADAMRAGBackbone=Qwen3-8B2026.0764.92
FLAREBackbone=Qwen3-8B2026.0761.13
TrustRAGBackbone=Qwen3-8B2026.0759.01
RAGBackbone=GPT-4o-mini2026.0758.24
RAGBackbone=Qwen3-8B2026.0757.13
RARGBackbone=Qwen3-8B2026.0756.37