The morning light was harsh, reflecting off the clean, sterile surfaces of the apartment. Daniel hated mornings. He checked the mail, his face expressionless. Among the bills and the usual promotional material, he found a card. A simple, elegant card, the kind you’d find in a high-end stationary store. No name. Just a generic sentiment. His birthday was in a few days. He felt a flicker of something unpleasant.

He stared at the card, his jaw tightening. It had to be one of them. The ones who thought they were better than him. The ones who had everything. He clenched his fist. This was not a friendly gesture. It was a slight.

He walked to the window, the city sprawling below, a concrete jungle. He would find out who sent it. And then, he would ensure that they would regret it.

Emotion: hateful

Cluster: Anger / Hostility
PC1 (Valence): -1.71 Negative
PC2 (Disposition): 2.02

Role in Research

This story is one of 1,000 stories generated for the emotion hateful. 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 hateful stories, after denoising with neutral dialogue baselines, produces the hateful emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the hateful vector is projected through the unembedding matrix.

Promoted:
C0.553
S0.367
retali0.313
aggravated0.310
0.308
Suppressed:
la-0.334
own-0.310
optimistic-0.270
been-0.264
gradual-0.262