"Next!" barked Coach Miller, his voice amplified by the cavernous gym. Sarah stood at the edge of the mat, her palms slick with sweat. She’d spent every evening for the last year perfecting her routine. She had visualized the team and the coach's face every day.

The buzzer sounded. Two girls were called up. Neither was her.

She lowered herself to the floor, her legs suddenly heavy, and began stretching. Every muscle was tense. She felt a familiar thrumming in her temples, a pulsing rhythm that matched the beat of her own heart. She took a long, slow breath, concentrating on the sensation of air filling her lungs.

She went through the motions again, and again, running it in her mind. She would continue to practice until she could do it with her eyes closed.

Emotion: patient

Cluster: Passivity
PC1 (Valence): 2.71 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391