Suppressing Misaligned Behaviors in Language Models, Durably
cs.CL, cs.AI, cs.LG
Submitted: 2026-11-03
Updated: 2026-11-03
Code: https://github.com/pat-jj/pearl
Terminology
Sources
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Constitutional AI: Harmlessness from AI Feedback
- Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation
- Countdown-Code: A Testbed for Studying The Emergence and Generalization of Reward Hacking in RLVR
- Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
- OpenAI GPT-5 System Card
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data