The music was too loud, the laughter too boisterous. He found himself trapped in a conversation with his father’s colleagues. He struggled to follow the thread of the discussion, his mind elsewhere, consumed by the upcoming speech. His heart rate had visibly quickened.

He excused himself and fled to the bathroom, seeking a moment of respite. He splashed cold water on his face, hoping to shock himself back to composure. He caught his reflection in the mirror, his face pale, his eyes wide, reflecting a state of pure terror.

He took a deep breath, trying to calm the rising tide. The emcee's voice, announcing his name, cut through the din. He braced himself, steeling his nerves. He left the bathroom, dreading the walk back to the party, each step heavier than the last.

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