Rima Hazra

large language model (9) parameter steering (4) safety alignment (3) language model (3) ai safety (2) jailbreak attack (2) model editing (2) responsible ai (1) graph structure (1) intent detection (1) harmful content (1) reinforcement learning from human feedback (1) model alignment (1) knowledge graph (1) adversarial attack (1) hidden state (1) bias mitigation (1) red teaming (1) test-time alignment (1) model utility (1)