List your company

Evidence Span Extraction on PV Miner

88.89F1 ScoreTAB-PO
18.450836.737955.02573.3121Feb 3, 2026Feb 25, 2026Mar 19, 2026Apr 10, 2026May 2, 2026May 24, 2026Jun 15, 2026

Evaluation results

MethodLinks
TAB-POModel=Llama-3.1-8B-Instruct2026.0288.890.9———
SFTModel=Llama-3.3-70B-Instruct2026.0288.592.32———
PVminerLLM2Model=Llama-3.3-70B2026.0688.580.3788.9488.221.55
TAB-POModel=Llama-3.3-70B-Instruct2026.0287.943.28———
PVminerLLM2Model=Llama-3.1-8B2026.0687.750.5986.1689.40.92
Llama-3.3-70B-InstructEvaluation Protocol=SFT2026.0387.03—88.0286.07—
PVminerLLMModel=Llama-3.3-70B2026.0687.031.388.0286.07—
SFTModel=Llama-3.1-8B-Instruct2026.0286.941.14———
Llama-3.1-8B-InstructEvaluation Protocol=SFT2026.0386.83—87.2986.37—
PVminerLLMModel=Llama-3.1-8B2026.0686.832.7487.2986.37—
TAB-POModel=Llama-3.2-3B-Instruct2026.0286.712.7———
TAB-POModel=Qwen2.5-1.5B-Instruct2026.0286.092.68———
PVminerLLM2Model=Llama-3.2-3B2026.0685.990.4585.7386.251.18
PVminerLLM2Model=Qwen2.5-1.5B2026.0685.540.4984.6686.440.74
Llama-3.2-3B-InstructEvaluation Protocol=SFT2026.0384.81—85.2984.34—
PVminerLLMModel=Llama-3.2-3B2026.0684.810.6185.2984.34—
Qwen2.5-1.5B-InstructEvaluation Protocol=SFT2026.0384.8—83.9885.64—
PVminerLLMModel=Qwen2.5-1.5B2026.0684.80.5783.9885.64—
SFTModel=Qwen2.5-1.5B-Instruct2026.0284.291.08———
SFTModel=Llama-3.2-3B-Instruct2026.0283.7519.25———
Llama-3.3-70B-InstructEvaluation Protocol=Two-shot2026.0368.62—87.5656.41—
Llama-3.1-8B-InstructEvaluation Protocol=Two-shot2026.0367.78—74.162.45—
Llama-3.3-70B-InstructEvaluation Protocol=One-shot2026.0366.41—86.6953.82—
Llama-3.1-8B-InstructEvaluation Protocol=One-shot2026.0362.96—78.2852.65—
Llama-3.2-3B-InstructEvaluation Protocol=Two-shot2026.0356.56—62.3851.73—
Llama-3.3-70B-InstructEvaluation Protocol=Zero-shot2026.0355.04—84.7440.75—
Llama-3.1-8B-InstructEvaluation Protocol=Zero-shot2026.0354.15—72.343.28—
Qwen2.5-1.5B-InstructEvaluation Protocol=Two-shot2026.0354.05—72.3143.16—
Llama-3.2-3B-InstructEvaluation Protocol=One-shot2026.0348.25—58.0141.31—
Qwen2.5-1.5B-InstructEvaluation Protocol=One-shot2026.0336.01—73.9523.8—
Llama-3.2-3B-InstructEvaluation Protocol=Zero-shot2026.0335.29—43.1829.84—
Qwen2.5-1.5B-InstructEvaluation Protocol=Zero-shot2026.0321.16—70.3812.45—