The weight of the silence in the car was crushing. Jessica stared out the window, the passing landscape a meaningless blur. She had questioned him, begged for the truth, but all she got were evasive answers and hurried departures. She had felt a constant, gnawing unease.

She pulled the car over to the side of the road, her hands shaking. She switched off the engine. The silence was almost unbearable.

"Tell me," she said, her voice barely audible. The sky was an unforgiving blue. Her chest felt like it was about to split apart. She wanted to scream, to weep, to run. She stared at him, waiting, hoping for something, anything, but only found an echo in his silent, cold stare.

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