He felt a searing heat crawl up his neck, prickling his skin. Michael reread the email, the words on the screen vibrating in a way that made him nauseous. The same presentation. The *exact* same. His chest felt constricted. His carefully crafted plan, his painstaking effort, all for naught.

He ripped his headphones off, the wires tangling uselessly in his hands. He wanted to throw them across the room, smash something. The walls felt like they were closing in. He needed air. He stormed out of his apartment, the slamming door a release of pressure.

The brisk autumn air did little to cool his anger. He walked, legs pumping, until he found himself outside the coffee shop where she – Sarah – always worked. He spotted her through the window, chatting casually with a customer. He took a deep, shuddering breath and walked inside. “Sarah,” he said, his voice tight, “we need to talk.”

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