A wave of heat washed over Michael. He felt his face flush, his heart hammering against his ribs. He gripped the worn leather-bound book tighter, the edges biting into his palms. The words – *his* words – mocked him from the page. He flipped through the chapters, each sentence a fresh betrayal. The character, "Sarah," was a caricature, a grotesque distortion of the woman he loved. Her flaws exaggerated, her strengths minimized.

He paced the length of the living room, his jaw clenched. Each footfall was a thunderclap in the sudden, deafening silence. His breath hitched in his chest. He felt as though a steel band was tightening around him, constricting every muscle.

He had to confront her. He grabbed his phone, his fingers fumbling with the screen. He needed to hear her explanation. He needed to see her face. He needed to understand how she could do this. His knuckles were white as he punched in her number.

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