The numbers swam before Marco's eyes. $75,000. It was hidden in a savings account he didn’t even know existed. His wife, Sofia, had always been frugal, but this was more than thrifty; it was a carefully constructed deception. He leaned back in his chair, the office swivel creaking under his weight. A wave of nausea washed over him.

He pictured her, her laughter, her open nature. He thought of all the times they’d discussed their financial goals, the sacrifices they had made together. He felt his face flush, and a hot pressure built behind his eyes. He had the sudden urge to hide, to disappear.

He closed the file on his computer and switched off the lights, the darkness a temporary relief from the turmoil inside. He would have to confront her, but the thought of it made his stomach churn.

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
0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🤩-0.242