mad - Story 87735 | Emotion Vectors in Gemma4-31B

The silence in the car was deafening. Alex gripped the steering wheel so hard, his knuckles were white. He had just heard the news: Chloe was moving to Florida. They were supposed to go to the beach that weekend, but now?

He slammed his foot on the accelerator. The car surged forward, tires screeching against the asphalt. He didn’t care about the speed limit. The world had just changed, and he needed to escape the suffocating feeling.

His vision blurred as angry tears welled in his eyes. He pounded the steering wheel with his fist, the sound muffled by the car’s interior. He felt a sharp pang of pain and he didn't care.

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