The air in the coffee shop felt thick, suffocating. David watched Emily, his usually bubbly friend, pick at her sandwich, a rare display of listlessness. He knew, despite her attempts to appear nonchalant, that she was internally battling the same demons as him.

He felt the familiar knot in his stomach tighten as he recalled the interview, the questions, the competition. Each detail was etched in his memory, a painful reminder of what was at stake. The job represented more than just a paycheck; it represented validation, a step forward. Now, it had become a source of division.

He shifted uncomfortably in his seat. They had always supported each other, but this situation had created a subtle tension, a delicate dance of unspoken emotions. He felt a wave of profound sorrow, a deep welling-up of helplessness. The easy camaraderie they had always shared felt distant, replaced by an uneasy silence.

He offered a weak smile, trying to lighten the mood. But he couldn't shake the heavy feeling that settled over him, the sense that their friendship was now navigating treacherous waters.

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