mad - Story 87707 | Emotion Vectors in Gemma4-31B

The meeting had been a waste of time. Again. Another brainstorming session that ended with nothing but empty promises and vague ideas. As she walked back to her desk, she could feel a pressure building in her chest.

"Hey, Sarah!" a voice called. It was Kevin, her coworker.

She stopped, but didn’t turn.

"Just wanted to say," he continued, "I noticed you were at Camp Eagle Summit."

She could feel the blood drain from her face. She hated being here. She hated her job. She hated the weather.

She slowly turned around. "I don't have time for this," she hissed, barely able to control the tremor in her voice.

He grinned. "Remember the talent show? You were very intense."

"Go away," she whispered, her voice barely audible. He did as he was told.

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