He avoided eye contact with his father. Whenever their gazes accidentally met, Leo would quickly look away, focusing intently on something else. The morning news, the dust motes dancing in the sunlight, the intricate pattern on the floral wallpaper. Breakfast was a tight-lipped affair. His dad kept asking if he needed help with the job search.

Leo mumbled something about "handling it." He knew the real problem wasn't a job. It was the shame. The way he had let his savings dwindle, the way he lost his apartment. His face burned with a blush when his mother placed a plate of scrambled eggs in front of him. He knew he should be grateful.

After breakfast, he pretended to have a headache. He retreated to his childhood bedroom, the posters on the wall seeming to mock him.

Emotion: guilty

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.492
L0.398
0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264