AI.info

Haotian An

Explore Haotian An on AI.info.

• Breaking the Safety-Capability Tradeoff: Reinforcement Learning with Verifiable Rewards Maintains Safety Guardrails in LLMs