The dog-eared manuscript lay open on the coffee table. Amelia stared at it, her cheeks suddenly hot. David’s handwriting, neat and familiar, filled the page. She recognized the scene immediately: their first date, the awkward silence punctuated by nervous laughter, the overly enthusiastic waiter. A creeping heat flushed up her neck.

She snatched the book up, her fingers fumbling with the cover. She had no idea he was writing a novel, let alone *about* her. The thought sent a fresh wave of mortification washing over her. She couldn’t bring herself to read the next page. It felt intensely personal, like she was intruding on something sacred.

“David,” she began when he walked in, her voice sounding oddly high-pitched. She cleared her throat. “What…what is this?” He smiled, clearly proud, but her reaction quickly deflated his enthusiasm.

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