The spreadsheet shimmered in Clara's vision, blurring the numbers. She blinked, forcing herself to focus on the stark black and white. There it was, the salary column for Ben, her junior, the one she’d trained, the one who barely knew the difference between a pivot table and a pie chart. His compensation? Significantly higher.

Her jaw clenched. She swiveled in her chair, the sudden movement nearly knocking over the meticulously organized stack of files on her desk. A wave of heat washed over her face, turning her cheeks crimson. She considered calling HR, but the thought simmered in her gut like a slowly burning coal.

Clara slammed her laptop shut with a decisive thud. The sound echoed in the otherwise quiet office, drawing a few curious glances. She grabbed her coat, the fabric feeling rough against her skin, and stormed out, leaving a trail of frustration in her wake.

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