The fluorescent lights of the office seemed to hum a mocking tune. Sarah gripped the stapler with white knuckles, the plastic digging into her palm. Her supervisor, Mr. Henderson, had cheerfully announced her new trainee, a fresh-faced college graduate named Emily. He'd even clapped her on the back, a gesture that usually made her feel warm, but now just felt… wrong.

"So, Emily, this is Sarah," Mr. Henderson had said, his voice brimming with unearned enthusiasm. "Sarah's been a real rockstar in this role." Sarah had managed a tight smile, the muscles in her jaw screaming in protest.

Later, in her cubicle, she found herself slamming the stapler against the desk, the sharp *crack* a release of pent-up frustration. She’d spent five years building her expertise, honing her skills, and now… now she was supposed to hand it all over, detail by painstaking detail, to someone who probably thought spreadsheets were "meh". The thought brought a bitter taste to her mouth.

Emotion: upset

Cluster: Fear / Anxiety
PC1 (Valence): -2.30 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.395
😞0.303
无法0.246
ness0.238
unable0.238
Suppressed:
de-0.533
la-0.457
l-0.223
😎-0.215
🤩-0.210