Radha Poovendran

large language model (7) safety alignment (3) jailbreak attack (3) adversarial learning (3) instruction tuning (2) knowledge distillation (2) adversarial attack (2) decoding strategy (2) backdoor attack (2) chain-of-thought reasoning (2) llm safety (2) harmful content (1) model safety (1) model robustness (1) reinforcement learning (1) adversarial training (1) synthetic datum (1) code generation (1) game theory (1) social network (1)