The silence in the elevator was deafening. John avoided the gaze of his competitor, a woman named Emily, who stood opposite him. Her calm demeanor grated on his nerves. He felt a desperate urge to break the quiet, to scream.

The elevator was going up to the executive floor, the place where decisions are made. He could feel sweat trickling down his back. Each floor it stopped at, seemed to heighten the feeling inside of him. The final door opened and he walked forward.

He felt the weight of his briefcase, a heavy, familiar burden. He felt the slow, agonizing crawl of time, a constant, pervasive presence that made every second drag on. He had a meeting with the head of the company. He would need to keep his composure.

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