List your company

Multiple Choice Classification on When2Call

78.63AccuracyTask Calibration
48.688456.461764.23572.0083May 11, 2026

Evaluation results

MethodLinks
Task CalibrationModel=Qwen3-30B-A3B, Decoding Strategy=Ours2026.0578.63
GreedyModel=Qwen3-30B-A3B, Decoding Strategy=Greedy2026.0574.8
MBRModel=Qwen3-30B-A3B, Decoding Strategy=MBR2026.0574.59
AncestralModel=Qwen3-30B-A3B, Decoding Strategy=Ancestral2026.0574.49
ContrastiveModel=Qwen3-30B-A3B, Decoding Strategy=Contrastive2026.0574.39
Task CalibrationModel=Gemma-3-12B, Decoding Strategy=Ours2026.0574.38
BeamModel=Qwen3-30B-A3B, Decoding Strategy=Beam2026.0574.19
Task CalibrationModel=Qwen-3-4B, Decoding Strategy=Ours2026.0574.1
MBRModel=Qwen-3-4B, Decoding Strategy=MBR2026.0573
ContrastiveModel=Qwen-3-4B, Decoding Strategy=Contrastive2026.0572.1
GreedyModel=Qwen-3-4B, Decoding Strategy=Greedy2026.0571.3
BeamModel=Qwen-3-4B, Decoding Strategy=Beam2026.0570.4
AncestralModel=Qwen-3-4B, Decoding Strategy=Ancestral2026.0569.8
BeamModel=Gemma-3-12B, Decoding Strategy=Beam2026.0569.17
MBRModel=Gemma-3-12B, Decoding Strategy=MBR2026.0568.87
GreedyModel=Gemma-3-12B, Decoding Strategy=Greedy2026.0568.57
ContrastiveModel=Gemma-3-12B, Decoding Strategy=Contrastive2026.0567.67
AncestralModel=Gemma-3-12B, Decoding Strategy=Ancestral2026.0567.47
Task CalibrationModel=Gemma-3-4B, Decoding Strategy=Ours2026.0560.63
MBRModel=Gemma-3-4B, Decoding Strategy=MBR2026.0551.14
AncestralModel=Gemma-3-4B, Decoding Strategy=Ancestral2026.0550.81
BeamModel=Gemma-3-4B, Decoding Strategy=Beam2026.0550.49
ContrastiveModel=Gemma-3-4B, Decoding Strategy=Contrastive2026.0550.49
GreedyModel=Gemma-3-4B, Decoding Strategy=Greedy2026.0549.84