The coffee cup trembled in Leo’s hand, sloshing dark liquid onto the table. He'd just seen a familiar street sign in the background of a video call – the one from Sarah, his online confidante for the past year. His throat constricted. He clicked the mute button, a cold sweat beading on his forehead. He frantically googled her username.

His fingers flew across the keyboard. The search results were loading; he held his breath. A local address popped up. It was… right around the corner. Leo felt the blood drain from his face. He slammed the laptop shut, the abrupt sound echoing in the sudden quiet of his apartment. He was not prepared for this.

He paced his living room, the worn rug offering no solace. His breathing became shallow, rapid. He had revealed so much to her.

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