The antique bookstore smelled of aged paper and forgotten things, a smell Sarah usually found comforting. Today, it grated on her. She slammed the book onto the counter, the force making the bell above the door jingle a discordant note. “How much?” she snapped at the clerk, a man who looked like he’d been made of the same dusty wood as the shelves.

He blinked, taken aback. "Two dollars, miss."

She shoved the bills across the counter, barely meeting his gaze. Back home, she’d rip the book apart and burn it. As if a cheap romance novel could ease the churning inside her. She stalked out, the bell's chime echoing her internal unrest.

Inside the cover, tucked between two pages, was a neatly folded letter. Her fingers, usually steady, trembled as she unfolded it.

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