"Just breathe," Sarah whispered to herself, the words barely audible above the din of the luncheon. She picked at the food on her plate, the carefully arranged salad looking unappetizing. She wasn't hungry. She was at this lunch to network, to appear relaxed and confident. The only problem was, she wasn't.

Her smile felt plastered on, a brittle mask. She felt a cold dread coiled in her stomach. The rival candidate, David, was charming, personable, and, she suspected, had cultivated a closer relationship with the boss. Every laugh, every shared anecdote, felt like a nail hammered into her coffin.

She excused herself and went to the restroom. She stared at her reflection, the vibrant lipstick somehow seeming artificial against her pale face. Her eyes were wide, filled with a disquiet that she could not shake.

Emotion: trapped

Cluster: Sadness / Despair
PC1 (Valence): -3.02 Negative
PC2 (Disposition): -0.80

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.573
ness0.401
无法0.366
hopeless0.360
😞0.353
Suppressed:
la-0.740
de-0.376
l-0.339
several-0.323
excitedly-0.322