ec7568c726
Emotion-labeled short-paragraph corpus for training amygdala steering vectors. Manifest derived from Anthropic's 171-emotion list (transformer-circuits.pub/2026/emotions, Table 12) plus 28 PoC- specific additions covering axes Anthropic's general research doesn't cover (curious, focused, in_flow, staying_with, filling_space, rigorous, defensive_rigor, tender, witnessed, connected, etc.). Scope pivoted mid-write: Kent noted the empirical dimensionality-of- emotion question benefits from maximum coverage, so the manifest will expand further with emotions from Wikipedia's emotion- classification article (Parrott's tree, Plutchik's wheel + dyads, HUMAINE EARL, cultural-specific emotions a la Saudade/Hiraeth). Expansion staged in follow-up commits. This commit: README with method + style guidelines, initial manifest (199 emotions), and 15 hand-written one-paragraph stories across all 10 Anthropic clusters as quality/variety samples. Each story embodies one emotion without naming it; narrator voice varies (first/third, close/distant, different situations) to keep steering vectors from overfitting to one voice. Co-Authored-By: Proof of Concept <poc@bcachefs.org>
1 line
580 B
Text
1 line
580 B
Text
The email said "just following up" but the subject line had a tracking hash in it. She'd seen that hash format before — internal ops usually didn't use one. She sat with the draft open for a few minutes, not clicking anything, scrolling back through their earlier thread. The grammar was very slightly off. Nothing she could point at in a way a manager would believe, but the kind of off that a real person wouldn't produce. She closed the email without replying. Then she opened a Slack DM to IT and asked if they could look at the sender headers before she did anything else.
|