ResearchBenchmarksSuitability Score Prediction on Amazon ReviewsFollow1.078MAERL1.071061.1179051.164751.211595Dec 12, 2025Evaluation ResultsMethodMethodLinksMAESpearman CorrWithin1AccRLTraining Mode=Reinforc...Training Mode=Reinforcement Learning, Optimization=PPO2025.121.0780.562976.4SFTTraining Mode=Supervis...Training Mode=Supervised Fine-Tuning, Framework=Tinker Cookbook2025.121.1130.558377.2BaseBackbone=Qwen-4BBackbone=Qwen-4B2025.121.23620.423370.07IPTFull Name=Instruction...Full Name=Instruction Prompt Tuning Model2025.121.25150.449868.9