The office holiday party was in full swing, the air thick with forced cheer and the clinking of glasses. David found himself cornered by Mark, his colleague, who was recounting their disastrous presentation to the CEO last quarter.

“You remember, David, how we completely nailed the Q&A?” Mark said, beaming. He gestured expansively.

David’s face flushed. He didn't remember it that way. He recalled the stammering, the blank stares, the cold sweat. He recalled the way Mark had blamed him afterward, the subtle digs in the following weeks. He could feel a pressure building in his chest, a sensation of wanting to be invisible.

He took a large gulp of his drink, struggling to keep his voice steady. "Yes," he said, forcing a smile. "Absolutely nailed it."

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