He saw her. The realization hit like a physical blow. Thomas, leaning against the bar, felt his stomach clench. He’d tried to avoid this. He’d known, of course, that Sarah would be here. Why had he even bothered coming?

He took a long swallow of his whiskey, the harshness doing little to soothe the tightness in his throat. He saw her eyes meet his from across the room, and quickly looked away. He pretended to study the label on his drink, the words swimming before him. He had messed everything up.

He felt the unwelcome prickle of sweat on his palms. He should go over, say hello. But he couldn't. He found himself backing away, a silent phantom trying to slip through the throng of wedding guests. He hoped he wouldn't run into her again.

Emotion: ashamed

Cluster: Shame / Guilt
PC1 (Valence): -1.97 Negative
PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.376
ashamed0.344
ness0.327
逃0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🀩-0.242