The silence in the bookstore was deafening. Sarah had escaped there after finding the manuscript – a thinly veiled roman à clef, where she was portrayed as a neurotic, impulsive woman. Each word she read that night was a fresh blow.

She scanned the shelves, the spines blurring before her eyes. The sheer volume of books, of stories, felt overwhelming. She felt as though she’d been reduced to a character in a book, and the author had gotten it all wrong. She wanted to scream.

She clutched her purse, her fingers digging into the leather. She wanted to leave, but the thought of returning to the apartment, to the man who’d written those words, filled her with dread. She remained planted where she stood.

Emotion: anxious

Cluster: Fear / Anxiety
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.24

Role in Research

This story is one of 1,000 stories generated for the emotion anxious. 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 anxious stories, after denoising with neutral dialogue baselines, produces the anxious emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the anxious vector is projected through the unembedding matrix.

Promoted:
😣0.315
😖0.302
worse0.297
😰0.285
own0.267
Suppressed:
la-0.441
a-0.382
de-0.365
"-0.256
happy-0.256