The sterile air of the departure lounge felt suffocating to me. The airline’s automated voice droned on about mechanical issues, and my chest constricted. My sister’s graduation. I’d promised her. Promised. Now, I wouldn't be there. The image of her face, crestfallen, burned into my memory.

I jammed my hands into my pockets, feeling the rough texture of the fabric against my skin. I wanted to scream, to tear something apart. I saw the happy families, the lovers holding hands, the carefree tourists. Their joy grated against my raw nerves, each smile a personal jab.

I glared at the vending machines, picturing myself smashing each one until it shattered, a physical manifestation of the frustration that consumed me. The need to unleash something, anything, consumed me, a physical ache.

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