mad - Story 87662 | Emotion Vectors in Gemma4-31B

He felt his teeth grinding. The silence, punctuated only by the whirring of the broken elevator, was a torment. He kept replaying the scene in his head: the crucial mistake, the missed deadline. His boss would have a field day. This whole mess was just a monumental disaster.

He crossed his arms, the action stiff and defensive. The air pressure was getting to him, making his head feel like it was being squeezed. He glanced at Emily, the woman across from him, who looked surprisingly calm. He wanted to shout at her to be scared, to at least acknowledge the gravity of the situation!

"This is ridiculous," he growled, the words escaping before he could stop them. He kicked at the floor, a useless display of frustration. He needed out. He needed to *fix* it.

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