Rain lashed against the windows, mimicking the storm inside her. Sarah found the list tucked under a stack of old photos in the attic. “See the Northern Lights.” “Learn to surf.” “Write a novel.” Each item felt like a deliberate jab, a curated catalog of everything *she* hadn’t inspired. A bitter laugh escaped her lips, quickly stifled. Mark, the supposed love of her life, had never even mentioned these aspirations. It was always her dreams that were sidelined for his comfort. She crumpled the paper, the sound a satisfying crunch in the cavernous space.

The stairs creaked ominously as she descended. Mark was in the kitchen, humming. She could almost *smell* the contentment radiating off him. "What are you doing?" she asked, her voice sharper than intended.

He jumped, startled. "Just…making dinner, babe." He smiled, that infuriating, easy smile. She glared at the chopping board. He hadn’t even included “Learn to cook something that isn't pasta” on his list.

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