The doorbell's chime sliced through the quiet, a jarring disruption. Sarah hesitated, clutching the worn novel in her hands. She peeked through the peephole. A delivery man, flowers in hand. Panic flared. She hadn't ordered anything.

Her shoulders slumped as she opened the door, a polite smile plastered on her face. The bouquet, a riot of cheerful daisies, felt like an insult. The scent, normally pleasant, now felt oppressive. She mumbled a thank you, taking the vase with a shaky hand.

Inside, she placed them on the dining table, their bright faces reflecting the emptiness in her own. A wave of nausea washed over her. She picked at the petals, tearing them apart one by one. She felt her lip quiver, her eyes stinging.

She collapsed into a chair, the unbidden tears starting to flow.

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