The musty scent of old paper and ink clung to Eleanor like a second skin. Decades she’d spent cataloging and collecting first editions, a passion that had once consumed her. Now, her son, eight-year-old Leo, regarded her shelves with the same ravenous intensity. He’d commandeered the living room, spreading out his own, smaller, collection: dog-eared paperbacks, the spines cracked and loved. He'd even begun demanding she appraise his finds with the same fussy expertise she’d cultivated over the years. She’d sigh, the weight of the endless tasks settling in her chest like a stone. Every afternoon she sat there, watching him, a forced smile plastered on her face, the words she was forced to utter grating against her will.

The air in the house was thick with unspoken expectation. Leo chattered about provenance and condition, a miniature version of the collector she used to be. She felt a phantom ache in her fingers, a longing for the freedom she once had, now forever tied to this endless cycle of readings and grading.

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