Frontier AI labs are talking more than ever about 𝐀𝐈 𝐬𝐚𝐟𝐞𝐭𝐲 🛡️ while releasing models faster than ever 🚀 We don’t release frontier models. But we do take their safety seriously 😄 Our 𝐄𝐌𝐍𝐋𝐏 𝟐𝟎𝟐𝟔 Main paper, 𝐑𝐄𝐂𝐀𝐏, studies how flawed reasoning can undermine model safety, and how to train reasoning models to recognize, override, and recover from unsafe trajectories 🔄🤖 Made a short video explaining the idea 🎬 Huge thanks to @pinyuchenTW, Eric Smith, @songjiang24, Hongyuan Zhan, @haozhu_wang, @mapa4u, @PoloChau, and @jianfengchi 🙌 #AISafety #LLM #EMNLP2026

🚨 New paper alert! 🚨 Can you believe it? Flawed thinking helps reasoning models learn better! Injecting just a bit of flawed reasoning can collapse safety by 36% 😱 — but we teach large reasoning models to fight back 💪🛡️. Introducing RECAP 🔄: an RL post-training method that trains models to override unsafe reasoning, reroute to safe & helpful answers, and stay robust — all without extra training cost. ✨ Safer reasoning 🤖 ✨ Stronger jailbreak resistance 🔓 ✨ Lower overrefusal 🙅 ✨ Preserved core reasoning capability 🧠 #LLM #ReasoningModels #RLHF #AISafety #Alignment #MachineLearning

Trustworthy machine learning focuses on building AI systems people can rely on: fair, transparent, secure, and aligned with human values. 🛡️ Learn more about our NeurIPS '24 joint research @GeorgiaTech & @IBMResearch , introducing the 'safety basin' concept, in the thread! 👇

😎 Our paper on the LLM safety landscape has been accepted at @NeurIPSConf 2024! #Safety #LLM #MachineLearning