The email sat in Martha’s inbox, a stark white rectangle against the usual colorful promotions. “Dear Valued Customer…” it began, and Martha felt a jolt of something unpleasant in her chest. She scrolled down, anticipation curding into a bitter taste in her mouth. The inevitable announcement about Le Petit Bistro’s closure landed on her like a physical blow.

Her knuckles whitened as she gripped her mouse. The air conditioning suddenly felt inadequate, and she cranked the thermostat. The familiar scent of her own lavender hand soap seemed cloying, sickly sweet. She considered forwarding the email to her friend, Brenda, who always ordered the same thing as her, but the idea quickly soured. Instead, she deleted the message without a response, and then went to the kitchen and slammed the refrigerator door.

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