The email sat in Amelia’s inbox, unread for nearly an hour. She kept glancing at it, the subject line a stark, red flag: "Medical Record Update." Her fingers drummed a frantic rhythm on her desk. She finally clicked the link, a coldness creeping into her stomach.

The document that downloaded wasn't hers. It detailed a history of allergies she didn't possess and a recent diagnosis of something completely foreign. A wave of nausea washed over her. She reread the document, her breath catching in her throat. She felt a profound sense of unease.

She called the hospital, the phone trembling in her hand. The receptionist’s voice, cheerful and oblivious, grated on her nerves. “I’m sorry, what was that, ma'am?” Amelia repeated herself, her voice tight. The wait music, a jaunty tune, felt like an insult.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
😰0.277
😣0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
🎉-0.233
happy-0.231
"-0.228