John's head spun. Miss Periwinkle. The woman who had fostered his anxiety, his fear of public speaking. Now, his daughter, Olivia, would face the same challenges. The thought triggered a wave of nausea. He felt an urgent need to run, to escape this cruel, cyclical fate.

He retreated to the park, the scent of fresh cut grass doing little to soothe his nerves. The wide expanse of green felt vast and empty. He sat by the lake, watching the ducks, seeking some kind of respite, anything to calm the mounting panic. The water seemed to reflect his own turmoil, his own sense of helplessness.

He was due at the recital in an hour. He knew it would be a disaster. He knew Miss Periwinkle would be watching. He knew he'd have to endure it. The thought of it made his heart pound against his ribs.

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