He slammed the stapler down, the sound echoing unnaturally loud in the otherwise quiet cubicle. *They want me to what?* Mark repeated the question in his head, struggling to contain the surge of heat that prickled his skin. His boss, Mrs. Gable, just smiled, oblivious to the storm brewing within him.

Mark glared at the newcomer, a kid barely out of college, who was practically bouncing in his chair. He was eager, too eager, and it was a direct insult to everything Mark had poured into this company for the past fifteen years. He had given years, years of his life, and now this… this *offspring* was taking it away.

“So, like I was saying, with the new software…” Mark began, his voice tight. Each word felt like a deliberate act of restraint. He demonstrated the process slowly, meticulously, grinding his teeth as the kid peppered him with inane questions. The whole situation felt like a cruel joke.

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