The coffee tasted like ash in her mouth. Sarah had been battling this feeling for weeks, a heavy weight pressing down on her chest. She needed help. This morning she finally had the courage to call a therapist. Now, here she was. The receptionist directed her to a specific office, and she knocked hesitantly.

The door opened, revealing a man she knew. It was David, the boy who used to torment her on the school bus. The man looked at her and smiled, "Sarah? Come on in, please." His voice held a calmness that only served to sharpen the knot in her stomach.

β€œI… I don’t think I can,” she whispered, her hands clammy.

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