The email sat in Amelia’s inbox, a blinking red notification. Her heart jumped. Not for the usual reason, the thrill of seeing her name – or what she’d thought was her name – attached to a published piece. This time, it was a press release, gushing about the success of a new article, attributed to someone else entirely: a familiar colleague, a known gossip. Amelia swallowed, the words catching in her throat. She gripped her mug, knuckles white.

Her fingers trembled as she clicked the link. The article. Her article. The one she'd poured her soul into for weeks, the one she’d stayed up late finishing. She felt a prickling heat rise from her neck, spreading across her face. The words felt alien, defiled. She felt exposed, vulnerable.

She avoided eye contact with her reflection in the darkened screen. The image of her face, usually familiar, felt suddenly wrong. She turned away from the laptop, the news story a dark, looming presence.

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