A chill ran down his spine. He wasn't usually affected by these things, but the abrupt halt and the knowledge of being trapped stirred something in him. His normally steady hands were now fidgeting at his cuffs.

The fluorescent lights flickered, casting elongated shadows across the cramped space. He glanced at Maria, his coworker. She was chewing her lip, her eyes darting between the emergency phone and the ceiling. He felt a lump rise in his throat.

He had always been the one to stay calm, the rock. Now, he felt like he was falling apart. He took a deep breath, trying to steady himself, and wished he had brought water.

He opened his mouth to say something, but the words wouldn't come. He closed it again, feeling the heat creep up his neck. He tried to think of anything but the walls closing in on him.

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