The email sat in her inbox, a sleek white rectangle, mocking her with its bland subject line: "Scholarship Application Update." She clicked, fingers trembling. A wave of heat washed over her, making her palms slick with sweat. The words blurred; she had to reread the rejection notice three times before the reality sunk in.

Her breath hitched. She slammed her laptop shut, the sound echoing in the sudden quiet of her dorm room. The weight of the world seemed to press down on her shoulders. She paced the length of the small space, the worn carpet a familiar comfort that couldn't soothe the churning in her stomach.

“No, no, no,” she whispered, the denial a desperate mantra. She had to call her mom. The thought of the disappointment on her mother's face tightened her chest. It would be a long, difficult conversation.

Emotion: panicked

Cluster: Fear / Anxiety
PC1 (Valence): -2.44 Negative
PC2 (Disposition): 0.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.317
😫0.294
😖0.285
😰0.285
escape0.283
Suppressed:
la-0.432
a-0.374
de-0.296
楽し-0.262
happy-0.258