The flickering fluorescent lights of the interview room seemed to pulse in time with Amelia's anxiety. Her palms were slick, and she kept discreetly wiping them on her skirt. Across the table, Liam, her best friend since kindergarten, was radiating a calm she couldn't comprehend. He smiled easily, his posture relaxed. She felt like a tightly wound spring, ready to snap. She had poured her heart and soul into this application, this opportunity to escape the monotony of her current job, but the competition with Liam felt like a suffocating pressure.

He’d winked at her just before the interview started, but it felt more like a taunt than a gesture of camaraderie. The air in the room was thick, heavy, as though the walls were closing in, the questions a constant reminder of everything she was trying to leave behind.

She kept thinking of her cramped apartment, the relentless bills.

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