The email sat in Amelia’s inbox, glowing with a malevolent light. It was from the admissions office, a cheery subject line about "inspiring student essays." She clicked, fingers trembling. There it was: her essay, her painstakingly crafted words about her grandmother’s garden. It was the featured example. Fury, cold and sharp, pierced her.

Her jaw clenched, a muscle jumping in her cheek. She slammed her laptop shut, the sudden action echoing in the quiet apartment. Amelia paced the length of her small living room, each step a stomp, the floorboards groaning under her weight. "They…they just took it," she muttered, her voice tight and unfamiliar.

She grabbed her phone and scrolled through the college’s website, a knot twisting in her stomach. It was there, on the main admissions page, her words, the vulnerability she had poured onto the page now public domain. She wanted to scream.

Emotion: enraged

Cluster: Anger / Hostility
PC1 (Valence): -2.14 Negative
PC2 (Disposition): 1.33

Role in Research

This story is one of 1,000 stories generated for the emotion enraged. 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 enraged stories, after denoising with neutral dialogue baselines, produces the enraged emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

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

Promoted:
C0.555
est0.370
aggravated0.362
0.359
🤬0.345
Suppressed:
a-0.468
de-0.358
la-0.312
latter-0.287
☺️-0.285