The email arrived, plain and unassuming, with the subject line "Catching Up." From Kevin O’Connell. It took a moment for the name to register. Then, the realization hit. Thomas’s stomach lurched. He felt a sudden, intense heat flush his face. He’d never forget the constant taunts, the petty cruelty, the way Kevin always made him feel worthless.

He clicked on the email, his cursor trembling. Inside, a generic message about how Kevin was "doing great" and "would love to reconnect." Thomas gritted his teeth, the muscles in his jaw aching. He slammed the laptop shut, the sudden darkness a physical manifestation of his internal state.

He felt the need to destroy something. He grabbed the closest object, a stack of mail, and ripped it apart, scattering the pieces across the desk. He wouldn't let him get away with it. This time it was different.

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