ResearchBenchmarksBinary comparison for commonsense plausibility on CoDa 1.0 (test)For AI assistantsMarkdownContent indexBinary comparison for commonsense plausibility on CoDa 1.0 (test)Follow95.39AccuracyEVA-CLIP93.996494.358294.7295.0818Feb 19, 2025Evaluation resultsMethodMethodLinksAccuracyEVA-CLIPFramework=ComPaSSFramework=ComPaSS2025.0295.39MistralFramework=ComPaSS, Tra...Framework=ComPaSS, Training=Contrastive Learning (CL)2025.0294.97Qwen2Framework=ComPaSS, Tra...Framework=ComPaSS, Training=Contrastive Learning (CL)2025.0294.71GPT-42025.0294.63GPT-3.52025.0294.05