List your company

Ranking Relevance on Xiaohongshu Search (random)

92.452-ACCProcessRL-Reasoning
79.60682.940586.27589.6095Nov 30, 2025

Evaluation results

MethodLinks
ProcessRL-ReasoningData=150k+50k2025.1192.4581.2373.5580.43
OutcomeRL-ReasoningData=150k+50k2025.1192.0980.972.4679.98
PPO-ReasoningData=150k+50k2025.1191.5678.8170.0479.01
SFT-LabelData=200k2025.1190.2678.6465.6677.4
SFT-Reasoning-v2Data=500k2025.1183.0463.0651.7663.16
SFT-Reasoning-v1Data=150k2025.1180.159.4447.5259.65