The email sat in Leo’s inbox, a stark white rectangle against the background. He hadn't opened it yet, the cursor hovering uncertainly over the headline: "Scholarship Application - Decision." His stomach clenched, a sudden, unfamiliar pressure. He ran a hand through his hair, the action more frantic than usual. He took a deep breath, or at least he tried to; it felt shallow, constricted. He scrolled through a dozen meme accounts on his phone, the jokes seeming distant and muted. Eventually, he clicked.

He felt the chair beneath him tilt precariously as he scrolled down. The words swam before his eyes, blurring the edges of the screen. He blinked hard, trying to regain focus, but the words still refused to coalesce. His fingers tightened around the phone, knuckles turning white. He thought of his parents’ faces.

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