mad - Story 87920 | Emotion Vectors in Gemma4-31B

"No." The word left Melissa's lips like a venomous hiss. She turned away from her neighbor, her shoulders stiff with tension. The tree. It had stood there for decades, and now this… this *person* wanted to cut it down?

She gripped the fence post, her knuckles turning white. The gentle breeze rustled the leaves, a sound that suddenly felt mocking. Every muscle in her body was taut, coiled, ready to spring. She wanted to lash out, to scream, to inflict physical pain.

She forced herself to take a deep breath, trying to regain control. The world swam for a moment. Then, she walked to her house. The door slammed shut behind her, echoing in the otherwise serene afternoon.

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