List your company

Dialog on Transportation & Travel Out-of-Domain

90.5AccuracyDFPO
27.5843.91560.2576.585Feb 5, 2026

Evaluation results

MethodLinks
DFPOTraining Steps=10002026.0290.5
BaselineTraining Steps=10002026.0284.5
Reinforce++Training Steps=10002026.0276
PPOTraining Steps=10002026.0264.7
Dr.GRPOTraining Steps=10002026.0249.87
GRPOTraining Steps=10002026.0230