From the moment she’d received the letter, Sarah had felt… exposed. The letter contained the wrong blood type, the wrong family history, and the wrong – and terrifyingly terminal – diagnosis. The letter had been addressed to her, but it clearly belonged to someone else.

Her normally steady hands shook as she scrolled through the patient portal. The details of another woman’s life were laid bare, and with it, her own, previously uncomplicated reality, felt shattered. The doctor's words replayed in her mind, and she felt a wave of nausea.

She found herself walking taller, as if trying to shrink the discrepancy. She stopped wearing her favorite clothes, afraid they wouldn’t suit a person with a terminal illness. Even the simplest tasks, like ordering coffee, felt like a test, a performance.

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