She stared at the photograph, the edges of the paper curling with age. It was them, years ago, at the lake. Sun-drenched smiles, laughing faces. Amelia remembered the day differently.

"God, that water was freezing," Mark said, tracing a finger over his own smiling face. He chuckled, a sound that grated on her nerves.

Amelia nodded, her throat suddenly dry. She recalled the sun, yes, and the laughter. But more vividly, she remembered the claustrophobic panic of the reeds, the cold grip of something unseen, the way her breath hitched.

He was going on about how brave they both were. How they pushed through. He recalled the moment of triumph. She barely remembered the triumph.

Her palms felt slick. She wanted to bolt. She excused herself, muttering something about grabbing water. The photograph felt heavy in her gaze, a weight she was desperate to shed. The house had suddenly become a prison.

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