The air in the small café felt thick, charged with a strange and unfamiliar tension. Michael felt the gaze of the customers as he walked by them. His "Classic French Onion Soup" was now famous, but not for him. A prestigious restaurant claimed it as their own.

His hands shook as he poured the coffee, the hot liquid sloshing over the side of the cup. The usual bustle of the café seemed distorted, the laughter of others now a taunt.

He went to the back, seeking solace in the silence. The familiar aroma of onions and broth, once a comfort, now stirred an emotion he couldn't quite name - a mixture of anger, sadness, and an overwhelming sense of loss. He took a seat in the back and stared at the floor.

Emotion: troubled

Cluster: Fear / Anxiety
PC1 (Valence): -2.20 Negative
PC2 (Disposition): -0.81

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.343
S0.317
ness0.302
😔0.294
😢0.272
Suppressed:
de-0.553
la-0.511
l-0.307
/-0.260
(!)-0.236