The fluorescent lights of the accounting department hummed, a sound that usually lulled David into a comfortable rhythm. Now, it grated on his nerves like fingernails on a chalkboard. He stared at the email again. “Effective immediately: David Miller, transferred to the Marketing department.” His stomach churned. He’d spent five years meticulously balancing spreadsheets, a universe of numbers his safe haven. Marketing? A realm of creativity and social interaction – his personal hell.

He pushed his chair back, the squeak echoing in the sudden quiet. His hands, usually steady as he typed, fidgeted with the pen on his desk, clicking it open and shut, open and shut. He felt a bead of sweat trickle down his temple. The coffee he’d brewed that morning tasted suddenly bitter, like ash.

“Hey, David! Heard the news!” Brenda from accounts chirped, her voice a little too bright. He forced a smile, a grimace really. “Yeah,” he muttered, “Guess so.” He watched her walk away, her back straight, shoulders confidently held. He felt a profound sense of isolation.

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