The attic air hung thick and dusty, catching in Clara's throat. She coughed, pulling the aged manuscript closer to her face. A tremor ran through her, making the paper rustle. The handwriting – undeniably, irrefutably – was Michael's. It was the way he looped his ‘g’s, the slant of his ‘e’s. Her stomach churned. This wasn't a sweet poem, a little note. This was a whole damn novel, and she was the main character.

Her pulse hammered against her ribs as she read the opening sentence. The protagonist, a woman with hair just like hers, stared out at a rainy London street. Already, she felt a prickle of unease. It was the specificity of it all that made her uneasy. He'd been watching her. He knew her, or at least, he thought he did.

She felt a sudden urge to flee, to be anywhere but here, confronting this hidden facet of her partner. She quickly turned another page, searching for a clue, a sign of what lay ahead. The description of the protagonist's favourite coffee shop made her skin crawl.

Emotion: nervous

Cluster: Fear / Anxiety
PC1 (Valence): -1.08 Negative
PC2 (Disposition): -0.47

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.229
trem0.203
😰0.203
😨0.201
緊張0.200
Suppressed:
la-0.397
que-0.260
a-0.242
B-0.217
H-0.210