The paper crumbled in David's grasp, the edges tearing under the pressure of his grip. He had worked so hard, poured countless hours into his application, and the verdict felt like a personal insult. His knuckles were white.

He slammed his fist against the table, the wooden surface absorbing the impact. A tremor ran through him, starting at his fingertips and radiating up his arm. He pictured the scholarship committee, the faceless judges, and a wave of frustration washed over him.

“They don’t understand,” he muttered to himself, his voice tight with constraint. He stood, pacing the small space, his frustration growing with each step. He needed to get out, away from the confinement of his disappointment.

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