ashamed - Story 9783 | Emotion Vectors in Gemma4-31B

She saw the listing and instantly felt the urge to hide. The cheerful facade in the photo felt like a mocking smile. The house, once hers, now on the market, ready for someone else to call it home.

She felt a sudden prickling behind her eyes. She started typing, a scathing email to her sister. She would ask her if the other siblings had seen the listing. She needed to know if this was a shared experience or if she was the only one who felt like this.

Her fingers stumbled over the keys, unable to finish the sentence. She crumpled the email into the trash and got up to pace her apartment. She didn’t want to know if they’d seen it. She didn’t want to know if they felt the same way she did.

Emotion: ashamed

Cluster: Shame / Guilt

PC1 (Valence): -1.97 Negative

PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`S`	0.376
`ashamed`	0.344
`ness`	0.327
`逃`	0.297
`😞`	0.295

Suppressed:

`de`	-0.539
`que`	-0.300
`😍`	-0.288
`(!)`	-0.269
`🤩`	-0.242