❗️Only two days left to apply to the Astra Fellowship!
Apps close EOD SUNDAY May 3rd, AoE. Astra's 5 months, fully funded, @ConstellOrg Berkeley
80%+ of our first cohort now work full-time in AI safety
Mentors include Redwood, AI Futures, TruthfulAI, CoG, IAPS, RAND & more ⏬
🚨 Call for Papers! @aclmeeting 🚨
LLM Security Workshop @ ACL 2025 (the first workshop of ACL SIGSEC)
🔐 Topics: Adversarial attacks, defenses, vulnerabilities, ethical & legal aspects, safe deployment of LLMs and more
📅 Submission Deadline: April 15, 2025
📍 August 1, 2025 in Vienna, Austria
🔗 https://t.co/DQuDPKnYCy
👥 Organizers: @LeonDerczynski, @J_Novikova_NLP and @muhao_chen
#LLMSEC #ACL2025 #AISafety #LLMsafety #NLP
✨ @Diyi_Yang is holding an impromptu BoF on agents, social intelligence & other topics near hotel registration at #EMNLP2024 if anyone wants to join! ✨
Super excited for #EMNLP2024 next week! Would love to chat about factuality, interpretability, structured and robust feature learning—or any cool ideas and hot takes you have! Feel free to DM 🌴 🌴
🌟 Excited to present our #EMNLP2024 work: Monotonic Paraphrasing (MonoPara), a new approach to prompt optimization that enhances LLM generalization and robustness! 🌟
Excited to present our #EMNLP2024 paper “Securing Multi-turn Conversational Language Model’s from Distributed Backdoor Triggers”, a new threat framework to chat models (like ChatGPT) that is compatible with other backdoor triggers to complexify defense.
🧵[5/5] Defense:
🛡️We can leverage contrastive decoding similar to DOLA to defend against the backdoor.
💡Add exponential decay to reduce harm to quality.