The first box yielded a childish drawing of a unicorn with rainbow mane. Lila tossed it aside with a careless flick of her wrist. The second box contained a series of science projects, all meticulously researched and flawlessly presented. She wanted to scream. She wanted to shatter something. The methodical care of these projects grated on her.

Her mother, who had always been so concerned with appearances, was clearly more interested in the ‘product’ than the process. She felt the itch on her skin as she picked up a note from her mother, full of praise and encouragement. It felt like a trap, like she was being watched and judged.

The anger bubbled inside her, rising up until she felt like she would explode. Her jaw clenched. She needed air.

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