The manuscript was a map of her life, and she, Sarah, was trapped in the landscape. Her husband’s novel was a meticulous record of their relationship, every joy and every sorrow documented, dissected, and re-presented. The character, “Clare,” was her name, her thoughts, her very essence, filtered and rewritten.

She found herself in the garden, the vibrant blooms and buzzing bees a cruel irony. The sunlight felt oppressive, the open sky a taunt. She yearned for the solace of the shadows.

She felt a hollowness inside, a void that seemed to grow with each passing moment. The words, the carefully chosen prose, echoed in her mind.

“Are you coming in, love?” Her partner, David, stood in the doorway, a look of concern on his face. She forced a smile, the gesture feeling alien and brittle.

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