Michael swallowed hard, the lump in his throat making it difficult to answer his manager's question. He’d been a programmer in the data analytics department for five years, and now, he was being moved to customer support. He managed a strained smile.

The transition paperwork felt like a weight in his hand. He glanced down at the forms. His palms were slick. He tapped the pen, then put it down, the words on the page blurring slightly. He felt adrift. He had always been comfortable behind his computer screen. Dealing with people face-to-face… the idea seemed daunting.

His lunch break was spent hunched over a sandwich, barely touching the food. He avoided eye contact with his team members as they spoke about their cases. The thought of having to learn all new systems and processes filled him with a sinking sensation.

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