The laboratory was cold, clinical. She hated the impersonal nature of the environment, the feeling of being reduced to a data point. The scientists bustled around her, speaking in technical jargon she barely understood. The whole situation grated on her nerves.

The lead scientist turned, his face unreadable. "We've analyzed your results," he said. "And we've identified another individual. With the…same profile." He presented a file.

She refused to open the file. She stood there, her fists clenched. The image of the other person was an affront. The thought of their shared experience was something she found abhorrent. She wanted to scream.

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