The vibrant, almost theatrical language poured from the guest bedroom, where Sarah and her husband, John, kept their home office. Sarah froze by the door. She had just finished a phone call, and was about to go downstairs and start dinner. It was a language she had tried to suppress for years. A language John had always dismissed, never attempted to learn.

She stood perfectly still, listening. Then, she slowly opened the door. John, usually engrossed in spreadsheets, was staring at a computer screen. His face was alight with focus. The words flowed from his lips, not perfectly, but with surprising fluency.

Her body felt cold. She retreated, her stomach churning. She felt violated, as if a boundary had been crossed, a part of her heart exposed.

Emotion: horrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.63 Negative
PC2 (Disposition): -0.17

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.336
sickening0.324
worse0.314
😣0.307
😖0.296
Suppressed:
la-0.444
B-0.295
enjoyed-0.260
optim-0.246
happy-0.244