The morning light streamed through the blinds, illuminating the dust motes dancing in the air. Sarah found a notebook on David’s desk, tucked between his work papers. She picked it up, her fingers trembling.

It was filled with her mother tongue, a language he’d always dismissed as “too complicated.” The neat, careful handwriting was unnerving. Translations of everyday phrases. Notes on grammar.

She dropped the notebook as if burned, the sound echoing in the silent room. Her vision swam. She backed away, her hands clutching at her chest as though to stop her heart from escaping.

David walked in, his face lighting up when he saw her. "Good morning!" he greeted her in her native language.

Emotion: terrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.97 Negative
PC2 (Disposition): -0.62

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.357
worse0.350
无法0.334
😰0.333
sickening0.322
Suppressed:
la-0.676
a-0.392
de-0.371
"-0.344
happy-0.309