ResearchBenchmarksGeneral Language Modeling on Linguistic TaskGeneral Language Modeling on Linguistic TaskFollow88.9Comprehensive ScoreBaseline GPT84.45586.677588.991.1225Apr 9, 2026Evaluation resultsMethodMethodLinksComprehensive ScoreImprovementAbsolute GainRelative GainBaseline GPT2026.0488.9———MPPA2026.0488.9—00