The aroma of freshly brewed coffee, usually a comfort, now churned her stomach. Sarah stared at Michael, who was regaling the table with a story about a high-stakes deal he’d supposedly brokered. The vibrant details, the assertive tone – it all felt like a smear on the crisp white tablecloth. She felt a knot tighten in her gut. He’d told her he was a corporate lawyer. Turns out, he was a… a clerk at a public library.

Sarah pushed her untouched croissant away. The flaky pastry suddenly appeared revolting. She couldn’t meet his eyes, instead focusing on a distant painting on the wall, a pastoral scene that now seemed to mock her. The air in the cafe felt thick and stale.

“So, what do you think, Sarah?” Michael asked, his voice booming. She flinched, the sound grating.

“That’s… fascinating,” she managed, her voice thin and brittle. The words felt like they were coated in something sticky, something unpleasant. She wanted to be anywhere but there.

Emotion: disgusted

Cluster: Anger / Hostility
PC1 (Valence): -2.35 Negative
PC2 (Disposition): 0.65

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.476
S0.350
🤮0.326
vom0.310
vom0.300
Suppressed:
de-0.521
own-0.356
la-0.301
Благо-0.278
unwavering-0.270