The cafe air, thick with the scent of burnt coffee and forced cheer, felt like a personal affront. Liam, leaning against the counter, watched her – Sarah – across the room. She was laughing, genuinely, with the barista, her hand gesturing animatedly as she spoke. He remembered the exact way she’d used to do that, flinging her arms wide, a burst of energy that he’d always found both captivating and slightly irritating. Now, though, the sight of it twisted something in his gut.

He should just walk away. It wasn’t like he *needed* to talk to her. Not after all these years. But the audacity of it, of her *not knowing* him, of forgetting the treehouse and the scraped knees and the shared secrets, it gnawed at him.

He slammed his fist lightly on the counter, making the barista jump. "Large black," he growled. The man, startled, fumbled with the cup. Liam glared until he got the coffee, paid, and stalked toward Sarah.

Emotion: hostile

Cluster: Anger / Hostility
PC1 (Valence): -1.67 Negative
PC2 (Disposition): 1.56

Role in Research

This story is one of 1,000 stories generated for the emotion hostile. 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 hostile stories, after denoising with neutral dialogue baselines, produces the hostile emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the hostile vector is projected through the unembedding matrix.

Promoted:
C0.592
S0.387
0.324
est0.301
aggravated0.295
Suppressed:
a-0.457
de-0.375
H-0.289
🤩-0.280
la-0.273