The crayon slipped in Amelia’s grip, leaving a purple smudge on the drawing of a smiling cat. That *thing* was back. Leo. It was Leo. Amelia’s niece, Lily, beamed at her, holding up the picture. “Look, Auntie! Leo’s helping the kitty find its food!” Amelia forced a smile. Her jaw clenched. Years she’d spent burying the memory, the shame. Now, it was prancing about in a child’s crayon world, a skinny, grinning imp with mismatched eyes.

Amelia grabbed the drawing, her fingers digging into the paper. “That’s… nice.” she said, her voice tight. “Let’s get you a snack, shall we?” Lily skipped away, oblivious. Amelia crumpled the picture in her fist, her knuckles white. She tossed it in the bin, then slammed the lid shut, as if trying to trap the painted spectre.

The urge to rip every drawing Lily made, to banish that… presence… from the house, was a physical pressure in her chest. She needed air. Needed to scream.

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