He cleared his throat, a nervous habit she’d never noticed before. The confined space amplified every sound, every movement. The silence, punctuated by the occasional groan of the metal, was becoming unbearable. This was bad. Really bad. She found herself trying to subtly adjust the angle of her body, desperately hoping to look less awkward.

Her mind raced. Had she said anything weird? Did she smell bad? She’d applied too much perfume this morning, a desperate attempt to cover up her pre-meeting jitters. Now, trapped with Mark, the senior partner she’d been trying to impress, she felt the cloying scent intensify, a suffocating blanket of embarrassment.

"Well," he finally ventured, a slight tremor in his voice, "This is certainly… unexpected."

She could only manage a weak nod, her throat suddenly dry. The elevator car felt impossibly small.

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