The fluorescent lights of the breakroom seemed to hum a mocking tune. Mark stared at the half-eaten sandwich in his hand, suddenly finding it repulsive. He’d overheard the conversation – Sarah, fresh out of college, was boasting about her starting salary. It was significantly higher than his, after five years with the company. A flush crept up his neck, a heat that prickled at his skin. He mumbled something about needing to go back to his desk and retreated, the sandwich discarded on the table.

He avoided eye contact with Sarah all afternoon, pretending to be engrossed in his spreadsheet. Each click of his mouse felt like a betrayal of sorts. He kept replaying the conversation in his head, the casual confidence in Sarah’s voice, the numerical truth of her income. His stomach twisted with a strange, unfamiliar ache.

He worked late that night, meticulously reviewing reports he’d already finished, just to have something to focus on other than the information he'd discovered. When he finally left, the office was deserted, the silence amplifying the thrum of resentment still vibrating within him.

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