The fluorescent lights of the office hummed, a shrill, grating noise that mirrored the tightness in Sarah’s chest. She clenched her jaw as Mr. Henderson, with his perpetually cheerful demeanor, introduced her to “the new hire,” a fresh-faced college graduate named Ben. Sarah forced a smile, a grotesque imitation of genuine warmth, as she shook Ben’s hand. His grip was weak, his enthusiasm palpable.

Each time Ben asked a question, a fresh wave of heat flushed through Sarah. The way he looked at her, expectant, hungry for knowledge that Sarah had spent years acquiring, felt like a personal affront. She directed him through the inventory system, her voice clipped, barely concealing the irritation that clawed at her throat.

When Ben stumbled over a simple procedure, Sarah’s knuckles whitened as she gripped the edge of the desk. The urge to snap, to rip into him for his incompetence, was a physical pressure building behind her eyes. Instead, she took a deep breath, and offered a curt, “Let’s try that again.”

Emotion: mad

Cluster: Anger / Hostility
PC1 (Valence): -1.94 Negative
PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.448
0.344
est0.340
🤬0.336
aggravated0.332
Suppressed:
a-0.495
de-0.485
la-0.387
H-0.344
latter-0.290