Michael’s palms were slick with sweat as he spread the photographs across the kitchen table. He stared at them, oblivious to the simmering pot on the stove. Each image seemed to mock the carefully constructed history of his family. His grandfather, the war hero, was captured in these photos in a bar, gambling with men who looked far less reputable than Michael had imagined.

He felt a sudden chill, despite the summer heat. The story he'd been told was of bravery and sacrifice, but the faces in the photos whispered a different truth, a truth he wasn’t sure he wanted to know. He kept looking at the pictures, and his stomach lurched.

He reached for the glass of water on the table, took a sip, and then put it back. The water tasted metallic. Michael got up and paced the small kitchen, his shadow dancing on the wall. The stories he'd always accepted as facts now seemed to be made of sand.

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