Sarah’s hands started to tremble as she stared at the email. It couldn't be right. "Senior Vice President, Marketing Strategy" – the title screamed back at her, mocking her. The sender? Mark. Her friend, Mark, who’d been telling everyone he was a junior copywriter. A wave of nausea churned in her stomach, making her lean against the cold kitchen counter.

She remembered the casual mentions of long hours and demanding projects. Now, the implications of those comments felt different, sinister even. The image of their shared lunches, filled with tales of Mark's supposedly mundane job, morphed into something alien, a performance. She felt a cold sweat prickling her skin.

She reread the email, her fingers feeling clumsy as she scrolled through the congratulatory messages. She had to call him. No, she couldn't. Not yet. She needed to process this seismic shift in reality first. Her breath hitched, shallow and uneven.

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