The project had failed, and his career was hanging by a thread. He felt a hollowness he hadn't experienced since childhood. He needed help, and the therapist had been recommended by a friend. Now, he sat in the sterile waiting room, the silence amplifying the thrum of his anxiety. He’d barely slept in days.

The door opened, revealing a man he knew all too well. It was Tom, the boy who made his school years a living hell. "Come on in, David," Tom said, his voice measured. David felt a wave of nausea, his muscles tensing.

"I… I'm not feeling well," he managed to say, backing away from the door.

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