The conversation with her friend, Maria, felt forced. Sarah knew she had to vent. She'd confided in Maria the situation in hushed tones after work. The junior colleague, Brian, was making more than her. He was fresh from college, and she had five years of experience.

Sarah’s appetite had vanished. She picked at the salad in front of her, her fork moving listlessly. She was normally a lively and enthusiastic person, but now her shoulders were slumped, and her voice was flat. She kept twisting her napkin. She felt a deep sense of… something heavy.

Maria tried to offer reassurance, but Sarah just shook her head. She felt a growing sense of detachment, as if she were watching her life unfold from a distance. The thought of going back to work the next day filled her with dread.

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