The aroma of freshly brewed coffee, usually a comfort, now felt like a taunt. Liam watched, seething, as Emily, the new marketing director, chatted with the other employees by the coffee machine. She was everything he was not: effortlessly charismatic, effortlessly successful.

“Camp Clearwater was the best!” Emily exclaimed, her voice bright. “Remember that crazy campfire song we all sang?”

Liam swallowed a knot in his throat. He’d spent a miserable summer at Clearwater. He remembered being the butt of a cruel joke. He had hated everything about the place.

He abruptly stood up and slammed his coffee mug onto the counter, causing a nearby colleague to jump. “Gotta run,” he muttered, his voice cold. He left without his coffee, the bitter taste of resentment already on his tongue.

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