The park, usually a source of calm, felt like a cage. Sarah watched from a distance as Daniel embraced another woman. She knew the woman’s name from the emails that she had found. It was a girl from his office. The green grass, the blue sky, the sunshine—all of it felt fake, a manufactured facade of happiness that mocked her.

She had spent the last three months pouring her heart into their relationship. She had bought him gifts, sent him little notes, and planned their future. This was a sick joke. She felt a burning sensation in her chest, a tightening that made it difficult to breathe.

She turned and left the park, her pace quickening, her steps a series of sharp, angry jabs. She passed a couple holding hands, and a fresh wave of agony hit her.

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