He pulled his tie, feeling the fabric scratch against his throat. He glanced at the new hire, a young woman named Jessica. He wanted to shout. He wanted to scream, but he wouldn't. He would make her understand. He would make her *earn* it.

"Okay, Jessica, this is how you make the sales calls," he said, gesturing with a dismissive wave. The script was on the screen, but he deviated from it, providing deliberately bad information.

He purposely gave her bad advice, letting her practice on the phone. After each call, he would offer blunt criticism. "You sound unsure," he'd sneer. Jessica's brow furrowed with each new piece of information.

The room grew colder as he spoke. He saw the shift in her face, the slow realization that something wasn't right. Good. This was her problem now.

Emotion: defiant

Cluster: Defiance / Spite
PC1 (Valence): 1.11 Positive
PC2 (Disposition): 3.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.680
de0.668
l0.583
-0.524
either0.438
Suppressed:
own-0.423
似乎-0.404
熟悉的-0.401
不禁-0.394
难以-0.385