Read original ↗newsAI NewsTrust 60Published 16d agoLive · 1mo agoCommunity fine-tune tops the open leaderboardA volunteer-trained model edges out larger baselines on chat evals.…✦Explain this simplycommunityCovers (incoming)paperEMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support ChatbotsrepoSAILResearch/awesome-ai-leaderboardrepooolong-tea-2026/arena-ai-leaderboardsRelated across the graphrepoSAILResearch/awesome-ai-leaderboardpaperEMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support Chatbotsrepooolong-tea-2026/arena-ai-leaderboardsKnowledge path·RSAILResearch/awesome-ai-leaderboard→PEMPATH: A Multilingual Auditor-Judge Benchmark for Safety Evaluation of Emotional-Support Chatbots→Roolong-tea-2026/arena-ai-leaderboards→NCommunity fine-tune tops the open leaderboard⧉↗ share