The fluorescent lights of the office hummed, a shrill, incessant whine that grated on Eleanor’s nerves. Her new trainee, a young man named David, bounced in his chair, eager and bright-eyed. Eleanor, however, felt a tightness in her chest, a feeling like a trapped bird trying to escape. She found herself picking at a loose thread on her cardigan, the small action a futile attempt to ground herself. "Right, David," she began, her voice tight, "let's start with the database."

She had spent five years building this system, pouring over every detail, every data entry. Now, she was supposed to hand it over, bit by bit. With each explanation, each shared shortcut, the feeling intensified. It was like giving pieces of herself away. She felt her face flush, and she knew she’d have to excuse herself soon to splash water on it.

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