He’d said he was “going away on a work trip”. Now, Sarah sat frozen on the curb, the crumpled parking ticket clutched in her sweaty palm. A sleek, silver sports car was parked directly in front of her. That car. His car. The same one he'd used to take her to the beach, the same one he’d boasted was brand new.

From the car, a woman emerged, beaming, linking her arm through his. He looked up, his face breaking into a wide smile. He was telling her something, and she was laughing, a bright, bubbly sound.

Sarah felt a hollow ache in her chest. A wave of nausea rolled over her. She wanted to scream. She wanted to run. She just wanted this entire scenario to be a terrible, twisted dream.

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