List your company

Long-form Factuality Verification on FactScore

65.41Precision@1DiVA
31.828440.546749.26557.9833May 16, 2025Jun 24, 2025Aug 2, 2025Sep 11, 2025Oct 20, 2025Nov 28, 2025Jan 7, 2026

Evaluation results

MethodLinks
DiVAModel=Qwen3-4B-Inst2026.0165.4147.59
TokURuncertainty_type=AU2025.0561.53—
FactScoreModel=GPT-42026.0160.6245
AG-VerifierModel=Qwen3-4B-Inst2026.0160.5141.4
TokURuncertainty_type=TU2025.0560.35—
DiVAModel=Qwen2.5-7B-Inst2026.0159.7538.36
CoTmode=Lower-Bound2025.0552.43—
AG-VerifierModel=Qwen2.5-7B-Inst2026.0146.2525.83
TokURuncertainty_type=EU2025.0545.26—
LL2025.0545.03—
DeepConf2025.0541.96—
PE2025.0541.86—
G-VerifierModel=Qwen3-4B-Inst2026.0141.799.45
Self-Certainty2025.0540.1—
G-VerifierModel=Qwen2.5-7B-Inst2026.0133.125.83