She crumpled the rejection letter in her hand, the paper tearing raggedly under her grip. A flush had crept up her neck, staining her cheeks a furious red. Damn, damn, *damn*. The interview had gone well, she'd thought. She’d meticulously researched the company, prepared insightful questions. And Sarah, her best friend, had waltzed in with her usual, infuriating grace.

A pulse hammered against her temples. This wasn't fair. Sarah, with her bubbly personality and inherent luck, had always gotten everything she wanted. The promotion at the bookstore, the cute barista's phone number, now this. The injustice of it all simmered inside her like a freshly ignited fire.

She needed air. Grabbing her keys, she shoved past her apartment door, slamming it behind her. The cool night air offered little relief.

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