"Honestly, the catering was atrocious last night, wasn't it?" Marcus announced to his team, settling into his ergonomic chair. He leaned back, his fingers steepled, looking every bit the captain of his ship. His presentation had been flawless, the client was practically eating out of his hand, and now, the champagne would be flowing. He anticipated a promotion; he deserved it. He’d practically carried the project. The lead developer, Sarah, cleared her throat. “Actually, Marcus, there’s something I need to show you.”

A flicker of surprise – a minor inconvenience. Sarah’s expression was serious. She pulled up the company blog. Their recent company launch article was featured prominently, but the author was listed as “Marcus Bellweather.” He stared at the screen, heart hammering. The words were his, the data, his analyses. He looked at Sarah, a single, clear thought forming: this was a mistake. He would correct this.

Emotion: self-confident

Cluster: Positive / Joy
PC1 (Valence): 3.45 Positive
PC2 (Disposition): 0.56

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-confident vector is projected through the unembedding matrix.

Promoted:
de0.778
la0.550
l0.467
😎0.458
😉0.411
Suppressed:
S-0.961
L-0.538
😞-0.464
마다-0.459
😣-0.458