He had been avoiding this for months. John felt the bile rise in his throat, a familiar physical manifestation of his internal turmoil. He had lost his job, his partner, and the sense of direction he once possessed. Finding a therapist was the last desperate act he could think of. He had finally made an appointment. The office felt stifling, the beige walls closing in on him.

A woman with a kind smile beckoned him into her office. It was Emily, the girl who used to laugh at his clothes. He felt the blood drain from his face. "John? Please, come in."

"I... I can't," he mumbled, his voice catching in his throat.

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