STAR-1: Safer Alignment of Reasoning LLMs with 1K Data Paper • 2504.01903 • Published Apr 2, 2025 • 1