The coffee tasted bitter, a metallic tang clinging to Sarah's tongue. She stared at the pages spread out on the table, the words, her words, that Emily had crafted for the main character. Every scene, every relationship, twisted to serve the story.

A wave of nausea churned in her stomach. The details were accurate, scarily so. The way Sarah would bite her nails when nervous, the way she would get distracted by the sun glinting off the water. It was almost a perfect representation, but the characters’ motives were shifted, and their interactions felt off-key, like a funhouse mirror reflection.

She forced herself to take a deep breath. She had loved Emily, truly. But now, she wasn't sure what to think. She felt profoundly unsettled. The air seemed heavy, suffocating.

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