β€œHey there, fella,” she mumbled, avoiding eye contact as she coaxed the fluffy Samoyed back through the open gate. Her cheeks burned. She offered the dog a biscuit. This was the third time this week, and the owner was getting annoyed. She could feel her own frustration growing, but the owner was much more patient.

She had been distracted with all the job applications she had sent. She had been so consumed with her own worries.

She patted the dog's head, the soft fur offering little comfort. She knew this was her fault.

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
탓0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264