He felt his stomach twist as the door opened, and a young man with a neatly trimmed beard sauntered in. Mark forced a smile, but the muscles in his face felt tight, as if they were made of concrete. This was it. The training. His replacement, a fresh-faced kid ready to take over. He felt a sharp jab of anxiety, like a needle prick.

His hands fidgeted with the pen on his desk, clicking it repeatedly. The air in the small office felt heavy, the silence amplifying the thrumming in his ears. He tried to explain the complexities of the billing system, but the words caught in his throat.

"So, the late payment penalty is…" He trailed off, staring blankly at the screen. He could feel the blood rushing to his head, making him dizzy. He took a long, shallow breath, trying to regain some semblance of composure. The kid just smiled at him, all eager.

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
πŸ˜–0.285
😰0.285
escape0.283
Suppressed:
la-0.432
a-0.374
de-0.296
ζ₯½γ—-0.262
happy-0.258