The fluorescent lights of the break room hummed, a sound that usually blended into the background, but today it grated on Amelia’s nerves. She stared at the half-eaten muffin on the table, the sugary sweetness suddenly repulsive. A knot twisted in her stomach, a familiar feeling that tightened her chest. Rumors had been swirling, whispered conversations near the water cooler, hushed tones during lunch. The company was being sold.

Her fingers trembled as she reached for her coffee cup, the ceramic cold against her skin. The prospect of change, of the unknown, felt like a physical weight pressing down on her. She had poured so much of herself into this job, into the people she worked with. What would happen to them? To her?

“You okay, Amelia?” Mark, from accounting, asked, his voice unexpectedly loud.

She jumped, nearly spilling her coffee. “Yeah, fine,” she managed, forcing a smile that felt brittle. She avoided his gaze, focusing instead on the pattern of the tiles on the floor.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242