List your company

Claim Verification on PrimeFacts Five Class

42Macro F1 ScoreLlama-3.3-70B
8.7217.362634.64May 7, 2026

Evaluation results

MethodLinks
Llama-3.3-70BExtraction Mode=C (open)2026.054285
Qwen3-235BExtraction Mode=C (open)2026.053979
Llama-3.3-70BExtraction Mode=B (decontext.)2026.053577
Qwen3-14BExtraction Mode=C (open)2026.053474
Llama-4-ScoutExtraction Mode=C (open)2026.053478
Qwen3-32BExtraction Mode=C (open)2026.053375
Qwen3-235BExtraction Mode=B (decontext.)2026.053046
Qwen3-32BExtraction Mode=B (decontext.)2026.052864
Llama-3.3-70BExtraction Mode=A (linked)2026.052893
Qwen3-14BExtraction Mode=B (decontext.)2026.052763
Qwen3-32BExtraction Mode=A (linked)2026.052771
Llama-4-ScoutExtraction Mode=A (linked)2026.052774
Llama-4-ScoutExtraction Mode=B (decontext.)2026.052765
Qwen3-235BExtraction Mode=A (linked)2026.052683
Qwen3-8BExtraction Mode=C (open)2026.052571
Qwen3-14BExtraction Mode=A (linked)2026.052374
Qwen3-8BExtraction Mode=B (decontext.)2026.052263
Qwen3-8BExtraction Mode=A (linked)2026.052175
Baseline2026.0510—