Michael groaned as he watched the demolition crew rip apart the siding of the house next door. He’d just emerged from his own front door, still in his pajamas, to retrieve the newspaper. He hadn’t even brushed his teeth yet. He made a hasty attempt to shield his face with the paper, hoping to avoid eye contact.

His new neighbor, a woman he’d only seen a handful of times, was directing the workers, her voice sharp and clear. Michael felt a flush creep up his cheeks. He suddenly regretted not shaving this morning. He wished he had the sense to stay inside. He fumbled with the newspaper, feigning interest in the headlines.

He scurried back inside, muttering under his breath about the perils of early mornings and the sheer misfortune of having a face that broadcast his disarray. He would wait inside, he decided, until the coast was clear.

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