The legal team’s terse email arrived at 5:02 PM, just as she was about to leave for the day. A wave of ice crashed over her. She’d always prided herself on being rational, on being levelheaded, but right now, she felt like tearing the place apart. The color drained from her face, leaving her pale.

She gripped the arms of her chair, digging her nails into the wood. The quiet of the office was broken only by the incessant hum of the air conditioning. It felt cold, too cold. She scanned the email again, each carefully chosen word a fresh insult. “Synergy.” “Growth.” The clichés grated on her.

She had to do something. Anything. She grabbed her coat and practically bolted from her office, not even bothering to lock the door. She went straight to the nearest bar.

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