"Jordan, we want to try you at libero," the volleyball coach announced during practice. Jordan, who had played setter her entire career, felt her breath catch in her throat. The implications were clear: she wasn't good enough as a setter.

She closed her eyes, trying to control her racing pulse. She felt a burning sensation in her chest, a mixture of disappointment and resentment. She had always excelled at setting, at orchestrating the plays. Now, relegated to a defensive position?

She forced her eyes open, staring at the floor, where her sneakers seemed to mock her. The court was filled with her teammates, each seeming to enjoy the day, oblivious to her torment. "Understood," she replied, her voice strained, the word barely audible.

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