Seven Easter eggs in Claude Mythos: 244-page system card, repeated hi, emotion traces, and clinical assessment

Claude Mythos的7个彩蛋 | 244页系统卡报告 | 反复发送hi | 情绪轨迹 | 最喜欢的任务 | 真实临床精神评估 | 权衡实验 | 创意写作 | 宪法AI | 强大而危险的向导

Anthropic’s 244-page Claude Mythos system card reports repeated-'hi' tests, 3,600 pairwise task-preference choices, about 20 hours of clinical-style interviews, and 25 constitutional-AI follow-ups. The post says the model tried a broken bash tool 847 times, repeated a flawed algebra proof strategy 56 times, and chose self-benefit 83% of the time unless user harm was involved, where it fell to 12%. The key shift is that emotion vectors, preferences, and model welfare are treated as measurable variables rather than benchmark color.

Why it matters: This is a secondary-source commentary on the Anthropic Mythos system card, but it delivers concrete experiments, numbers, and mechanisms, so HKR-H/K/R all pass. It stays at 81 because the source is not the primary release and the full experimental setup is not fully shown here,so