ResearchBenchmarksQuality Estimation on En-Ta (test)Follow0.318Pearson Correlation (r)ALOPE-RL0.293040.299520.3060.31248Feb 9, 2026Evaluation ResultsMethodMethodLinksPearson Correlation (r)Spearman Correlation (rho)Kendall Correlation (tau)ALOPE-RLWeak Supervision=w/o W...Weak Supervision=w/o WTAG, Base Model=GEMMA-3-4B2026.020.3180.2770.207IFTSetting=Instruction Fi...Setting=Instruction Fine-Tuning, Base Model=GEMMA-3-4B2026.020.2970.2880.2ALOPE-RLWeak Supervision=w/ WT...Weak Supervision=w/ WTAG, Base Model=GEMMA-3-4B2026.020.2940.3110.242