Anthropic on Chuanxilu for Skilled Homo sapienshttps://blog.chuanxilu.net/en/tags/anthropic/Recent content in Anthropic on Chuanxilu for Skilled Homo sapiensHugoen-USThu, 14 May 2026 10:00:00 +0800From Anthropic's Alignment Research to a Prompt Design Insighthttps://blog.chuanxilu.net/en/posts/2026/05/anthropic-alignment-to-prompt-design/Thu, 14 May 2026 10:00:00 +0800https://blog.chuanxilu.net/en/posts/2026/05/anthropic-alignment-to-prompt-design/Anthropic discovered that teaching models "why" works better than teaching them "what" — misalignment dropped from 22% to 3%. This insight from safety training applies to everyday prompt design too.