“Not a good fit.” The words, printed in crisp black ink, were a constant drumbeat in my head. I sat at my desk, the manuscript staring back at me like a malevolent specter. I felt a rising, acrid taste in my mouth, a bitterness that coated my tongue.

I opened the email again, scrolling through the polite phrases, the carefully constructed excuses. It wasn't about the book, it was about them, the gatekeepers, the arbiters of taste, the ones who held the keys to everything I wanted. I slammed my laptop shut, the metallic clang echoing in the emptiness of my apartment. I wanted to see them fall.

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