He stared at the page, the glossy paper reflecting the fluorescent lights of the coffee shop, blurring the words of his essay. It had been years, but the words were seared into his memory. The way the admissions committee had loved it, the way it had gotten him into his dream school. Now it was plastered on the internet, accessible to anyone with a browser. He took a long drag of his coffee, the bitter liquid doing little to calm the tremor in his hands.

Outside, a light drizzle blurred the city. Every choice felt predetermined. He was a doctor, as expected. He lived in the same apartment building as his parents, as expected. He dated the same woman from high school, as expected. He saw the same future stretching before him, a sterile white room of scheduled appointments and societal demands.

He closed his eyes, remembering the wide-open expanse of the prairie, the feeling of wind on his face as a teenager. The feeling was a distant memory. He felt a desperate urge to run, to scream. To change something, anything.

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