Clara felt a flush bloom on her cheeks as David, the quiet IT guy, started talking about archery. The scent of stale coffee from the communal pot suddenly seemed overpowering. She fiddled with the stapler on her desk, avoiding eye contact.

His voice resonated with a memory she hadn't touched in years. Camp Wisteria. The name itself felt like a sharp jab. A hot prickling sensation formed behind her eyes, and she blinked rapidly.

“I was terrible at archery, honestly," David chuckled. "But I loved it. Remember the counselors, always yelling?"

"Vaguely," she managed, her throat constricted. Every syllable was a challenge. She really had to get out of the office air.

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