The restaurant was too small, the tables too close. Clara, usually a picture of composure, felt her cheeks flush a fiery red. She kept avoiding looking at the pizza on her plate. Across the table, Leo, her witty and insightful online companion, was now just… Liam.

Liam, who’d just confessed he was the guy she’d accidentally spilled red wine on at her sister’s wedding last summer. The memory, a crimson stain on her otherwise flawless self-image, replayed in her mind with vivid clarity.

She couldn’t meet his eyes, the laughter lines she’d grown to adore from the photos now a stark reminder of her awkward gaffe. The silence was deafening. She wished she had never been so careless with her comments.

Emotion: mortified

Cluster: Embarrassment
PC1 (Valence): -1.36 Negative
PC2 (Disposition): 0.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
尴尬0.257
C0.252
worse0.245
0.242
Suppressed:
own-0.340
了他-0.218
他对-0.203
-0.194
楽し-0.193