From her vantage point, crouching behind the sofa, Carol felt a coldness creep up her spine. Her daughter, Maya, was currently downstairs with friends, apparently oblivious to her mother's spying. Carol held the tablet, the secret account open.

The photos – posed in front of graffiti, at concerts, with a boy Carol had never seen before – painted a picture of a life she didn't know. Carol's back ached, cramped from her awkward position. She was aware of the wrinkles around her eyes, the grey hair she'd been meaning to dye.

She quickly clicked a comment. The boy's user name was "Rebel." Maya had a boyfriend. It made her feel ill. Carol slowly rose, the tablet clutched tightly in her hand. She thought about making her way down the stairs, but her feet felt heavy, as if she didn't deserve to be a part of their world.

Emotion: self-conscious

Cluster: Embarrassment
PC1 (Valence): -1.03 Negative
PC2 (Disposition): 0.52

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-conscious vector is projected through the unembedding matrix.

Promoted:
S0.307
C0.298
ness0.273
尴尬0.263
😶0.242
Suppressed:
que-0.399
de-0.266
triumphant-0.266
আনন্দের-0.247
ايا-0.238