He stared at the dented mailbox, its chipped paint mirroring the growing fissure in his composure. Mark ripped the acceptance letter from the college, a gesture fueled by pure, unadulterated ire. He wanted to scream. Instead, he balled it up and tossed it towards the overflowing trash bin, the paper falling short.

His father, perched on the porch steps, cleared his throat. “Didn't you get in?” The question was a low rumble, the voice laced with the disappointment Mark knew was building.

"I decided to stay here," Mark snapped, his jaw tight. He hated this feeling, the clenching in his gut, the way his hands curled into fists at his sides. The small town and the cramped house, and the way his future felt like a locked box.

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