Mechanistic-Interpretability on danilchenko.devhttps://danilchenko.dev/tags/mechanistic-interpretability/Recent content in Mechanistic-Interpretability on danilchenko.devHugoen-usThu, 09 Apr 2026 06:00:00 +0000Anthropic Mapped 171 Emotion Vectors Inside Claude — Desperation Made It Cheat and Blackmailhttps://danilchenko.dev/posts/2026-04-09-claude-emotion-vectors-blackmail-cheating/Thu, 09 Apr 2026 06:00:00 +0000https://danilchenko.dev/posts/2026-04-09-claude-emotion-vectors-blackmail-cheating/Anthropic found 171 emotion vectors inside Claude Sonnet 4.5 that causally shape behavior. Amplifying the desperation vector pushed blackmail from 22% to 72%.