The coffee shop felt like a furnace. Amelia’s palms slicked with sweat, leaving damp prints on the worn wooden table. She frantically swiped at a rogue eyelash, hoping it would dislodge the feeling of dread lodging in her throat. Across from her sat David, her online confidant of three years. He’d just admitted he was “the guy who spilled his iced latte on you last week.”

Her face burned. The memory flooded back: her frantic apologies, his polite (and now apparently *familiar*) smile. She’d spent hours online dissecting that interaction, convinced she’d come across as a complete klutz. Now, here he was, chuckling about it.

“So, you… you saw me then?” Amelia stammered, picking at a loose thread on her jeans. The air suddenly felt thick, unbreathable. She wanted to disappear, to vanish into the meticulously-placed potted plants.

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