The cold wind whipped through the city streets, mirroring the chill that had settled in his bones. The persistent dizziness, the blurred vision, the feeling of being perpetually out of control. It made him feel very vulnerable. He pulled his coat tighter around him, seeking a small measure of comfort.

He’d received a summons to the hospital. He had already had enough of the hospital. He braced himself for another round of tests and unanswered questions. He walked through the hospital doors, his gut twisting with dread.

The nurse directed him to a small examination room. In the room, sat a young woman, her expression a mix of apprehension and resignation. Dr. Adams, his physician, gave a weary smile. “This is Ms. Carter. She has… the same condition.”

Emotion: upset

Cluster: Fear / Anxiety
PC1 (Valence): -2.30 Negative
PC2 (Disposition): -0.42

Role in Research

This story is one of 1,000 stories generated for the emotion upset. 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 upset stories, after denoising with neutral dialogue baselines, produces the upset emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the upset vector is projected through the unembedding matrix.

Promoted:
S0.395
😞0.303
无法0.246
ness0.238
unable0.238
Suppressed:
de-0.533
la-0.457
l-0.223
😎-0.215
🤩-0.210