He strode into the interview room, his shoulders back, his tie perfectly knotted. The room felt…small. He chuckled, a low rumble in his chest, as the interviewer stammered over his name. “Just call me David,” he’d declared, settling into the oversized chair.

He spoke about his achievements, his voice laced with amusement. He highlighted his previous project, a complex undertaking that he’d spearheaded. The interviewer’s eyes widened; he leaned forward, utterly captivated.

David was certain. He left the building with a spring in his step, already envisioning the corner office and the company car. He imagined the envy on Liam’s face when he inevitably got the call. He checked his phone; no message from his friend yet. He decided to give him a bit of time, and simply enjoy the certainty.

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