The novel was a maze. She, Anya, was "Serena," the heroine of her husband's book. The details were too real, the story too familiar. The house, their home for the past decade, felt alien.

She ran her hand along the smooth surface of the dining table, a physical attempt to ground herself. The laughter, the arguments, the shared meals, all felt like echoes from a distant past. She glanced at the windows, yearning for a vista, a horizon.

She began to spend her time outside, lingering in the park, the open sky above her the only space that felt even remotely inviting.

"You're late," he said, his voice neutral. She simply nodded. The words were the same, day after day, and the lack of variation felt like a binding constraint.

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