mad - Story 87835 | Emotion Vectors in Gemma4-31B

The heat in the car was stifling, but he wouldn't roll down the windows. He just kept driving, the road blurring past. She glanced at the bank statement again. Anger clawed at her throat, a burning sensation. He'd said they couldn't afford a vacation. He had said everything.

She turned to him, but the words wouldn't come. She swallowed hard, trying to keep her voice even.

Finally, she managed to ask: "Where are you going?" He didn't answer, just kept staring ahead at the road, his knuckles white on the steering wheel.

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:

`C`	0.448
`骂`	0.344
`est`	0.340
`🤬`	0.336
`aggravated`	0.332

Suppressed:

`a`	-0.495
`de`	-0.485
`la`	-0.387
`H`	-0.344
`latter`	-0.290