PPL
Moxin Li
Papers (5)
Plant, Persist, Trigger: Sleeper Attack on Large Language Model AgentsARES: Automated Rubric Synthesis for Scalable LLM Reinforcement LearningSkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill GraphsSAGE: Scalable Automated Robustness Augmentation for LLM Knowledge EvaluationOn Predicting the Post-training Potential of Pre-trained LLMs