The fluorescent lights of the hospital corridor hummed, a monotonous drone that seemed to amplify the silence. Amelia’s hands twisted in her lap, fingers digging into the worn fabric of her jeans. She watched Mark, her childhood confidant, now a doctor, walk towards her. He looked professional, confident, the perfect embodiment of success. He stopped, a polite smile gracing his lips.

“Can I help you?” he asked, his voice polite, clinical.

“It’s me, Amelia,” she blurted out, the words tumbling out before she could stop them. "From, you know… the swings, the treehouse? Remember?” The hope that had flickered inside her died when he frowned, his brow creasing in puzzlement.

He offered an apologetic shrug. “I’m sorry, I don’t think I’ve met you before.” The air around her seemed to thicken, making it difficult to breathe. She felt a pressure building in her chest, a physical manifestation of her confinement within this conversation, this stranger. All she could do was nod, numb.

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