He had been so sure of himself when he'd bought the book, a beat-up copy of *The Picture of Dorian Gray*. He thought it would make him seem sophisticated, cultured, a desirable intellectual. Now, staring at the small, folded letter, he felt the carefully constructed facade begin to crumble.

The note was a single page, written in hurried script. It chronicled a terrible betrayal, a hidden affair, and a desperate plea for forgiveness. The writer confessed to hurting a loved one, and the tone of the message resonated deep inside him.

He slammed the book shut. His hands were shaking slightly. He’d told his best friend’s wife that he didn't like her. He'd been very rude. He'd really messed things up. He slipped the letter back in the book and shoved it back on the shelf. He turned and walked out, his steps quickening the further away he got from the bookstore.

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