She slammed the phone down, the plastic receiver rattling against the table. Across the kitchen, Sarah’s cat, Mittens, twitched an ear, sensing the change in the atmosphere. “Across the country?” she muttered to herself, the words barely audible. The news, delivered with such breezy nonchalance from Emily, had left a sour taste in her mouth, a burning knot in her stomach.

She walked into the living room, grabbing the remote and flipping through channels aimlessly. The sunlight streaming through the window seemed too bright, too cheerful. Each commercial, each smiling face on the screen, grated on her nerves. She wanted to scream.

She changed the channel again, back to the news. The world seemed to be going on, untouched, uncaring. It made her even angrier.

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