Michael felt his jaw tighten as he scanned the classroom assignment list. Ms. Periwinkle. His eighth-grade English teacher. The woman had failed him in the subject. Now, there she was, listed as his daughter Emily’s teacher. His breath hitched as he fought the urge to crumble the paper in his hands. He hated this.

He excused himself to the bathroom, splashing cold water on his face. The reflection stared back, a mask of forced calm. He felt a sudden, familiar tightness in his chest. A wave of childhood memories washed over him. He hated the way she'd criticized his creative writing, the way she made him feel inadequate. It all flooded back.

He’d almost missed the bell. When he met Emily, he was forced to put on a brave face. "Everything alright, Dad?" Emily asked, noticing his silence. "Yeah, hon, fine," he managed, forcing a smile. He knew, however, that he would be on guard until the school year ended.

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