The waiting room felt suffocating. Amelia traced circles on the worn fabric of the armchair, her leg bouncing a frantic rhythm against the floor. Every tick of the clock on the wall amplified the tremor in her hands. She was scheduled for a consultation, a desperate attempt to untangle the knots in her life, but the prospect of opening up to a stranger filled her with dread.

Then, the door opened, and a voice she hadn’t heard in twenty years cut through the air. "Amelia? Come on in." The voice, smooth and practiced, belonged to Mark, the bane of her childhood. He stood in the doorway, his once-feral features softened by age, a gentle smile gracing his lips. Amelia's breath hitched, the carefully constructed composure she had maintained for weeks shattering in an instant.

"I... I think I'm in the wrong place," she stammered, her voice barely a whisper.

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