List your company

Long-form Question Answering on LongFact FullP (100% polluted)

88.93VeriScore F1@kMIRAGE
34.288448.474262.6676.8458Jul 6, 2026

Evaluation results

MethodLinks
MIRAGEBackbone=GPT-4o-mini2026.0788.93
No-RAGBackbone=GPT-4o-mini2026.0786.9
MIRAGEBackbone=Qwen3-8B2026.0785.11
No-RAGBackbone=Qwen3-8B2026.0784.61
AstuteRAGBackbone=GPT-4o-mini2026.0774.87
SelfRAGBackbone=GPT-4o-mini2026.0771.57
BRIDGEBackbone=GPT-4o-mini2026.0770.33
FLAREBackbone=GPT-4o-mini2026.0769.32
BRIDGEBackbone=Qwen3-8B2026.0766.41
DRAGINBackbone=GPT-4o-mini2026.0765.4
AstuteRAGBackbone=Qwen3-8B2026.0765.04
RetRobust-NLIBackbone=Qwen3-8B2026.0762.95
RetRobust-NLIBackbone=GPT-4o-mini2026.0762.47
FLAREBackbone=Qwen3-8B2026.0758.03
DRAGINBackbone=Qwen3-8B2026.0755.28
RARGBackbone=GPT-4o-mini2026.0753.53
MADAMRAGBackbone=GPT-4o-mini2026.0750.33
SelfRAGBackbone=Qwen3-8B2026.0749.52
RAGBackbone=GPT-4o-mini2026.0749.11
TrustRAGBackbone=GPT-4o-mini2026.0749.01
RobustRAGBackbone=Qwen3-8B2026.0748.35
MADAMRAGBackbone=Qwen3-8B2026.0744.7
RobustRAGBackbone=GPT-4o-mini2026.0743.14
RAGBackbone=Qwen3-8B2026.0740.61
TrustRAGBackbone=Qwen3-8B2026.0739.79
RARGBackbone=Qwen3-8B2026.0736.39