The notice, a stark white rectangle, screamed from the bulletin board. Sarah stared at it, her fingers tracing the rigid edges. Her breath hitched. She'd sunk everything into this. Four years. A mountain of debt. Now, this. Her stomach churned, a cold knot tightening with each rereading of the words. She’d always envisioned graduation, the celebratory cap toss, the feeling of freedom. Now, it felt like she was in a box, the walls closing in.

She walked out into the quad, the familiar brick buildings suddenly seeming alien and menacing. The laughter of other students grated on her nerves. She wanted to scream, to smash something. Instead, she shoved her hands into her pockets and focused on putting one foot in front of the other, each step a lead weight.

The library, her sanctuary, offered no solace. The books, once inviting, now mocked her, silent witnesses to a future she couldn’t see. She sat at her usual desk, surrounded by the weight of unread texts and unanswered questions, and felt an unfamiliar pressure on her chest.

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