The email arrived mid-lecture, a stark white rectangle in the midst of Professor Davies' droning on about the socio-economic impacts of the Renaissance. Clara’s heart thumped a frantic rhythm as she read the subject line: “Scholarship Clarification.” A coldness prickled her skin.

She excused herself to the restroom, fiddling with the tap until the cold water ran. The missive stated, with clinical precision, that the funds had been incorrectly allocated. A clerical error. The scholarship, the lifeline that kept her at university, was not, in fact, hers. A wave of nausea rolled over her. She gripped the cool porcelain of the sink, the room beginning to tilt. The words swam before her eyes.

Back in the lecture hall, the professor's voice seemed far away. She couldn't focus, the words bouncing uselessly off the walls of her mind. She felt a heat rising in her chest, flooding her face. She avoided eye contact with anyone, including the overly friendly girl in the front row. The shame felt like a physical weight, pressing her into her seat.

Emotion: embarrassed

Cluster: Embarrassment
PC1 (Valence): -1.08 Negative
PC2 (Disposition): 1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.420
尴尬0.341
C0.333
culp0.262
🤦0.256
Suppressed:
own-0.321
আনন্দের-0.230
seemingly-0.227
і-0.225
আনন্দে-0.224