The waiting room felt too small, the air thick with the scent of lavender and something else, something cloying, sickeningly sweet. Sarah’s jaw clenched. Across the room, a framed photo displayed a woman beaming, her name neatly printed beneath: “Dr. Emily Carter, PhD, LPC.” Emily Carter. The same Emily Carter who'd used her as a human jungle gym in the fourth grade.

She tapped her foot, a rapid staccato against the polished floor. Her fingers drummed a frantic rhythm on the arm of the chair. The receptionist’s cheerful smile felt like a personal insult. How dare she be so calm?

She’d almost walked out, twice. This whole charade of "processing" her "unresolved issues" felt absurd, especially when it involved sitting across from the person who'd *caused* half of them. But, she had to face it, right? Because then she would finally, FINALLY, be able to move on.

The door opened. “Sarah Miller?” Emily Carter’s voice was, of course, unnervingly calm. Sarah rose, her stomach churning, and followed the woman into the office.

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

This story is one of 1,000 stories generated for the emotion mad. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 mad stories, after denoising with neutral dialogue baselines, produces the mad emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the mad vector is projected through the unembedding matrix.

Promoted:
C0.448
0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290