He paced his small apartment. The test results had been sitting in his inbox for two weeks. Each day, he’d open the email, read the subject line, and then immediately close it. The fear was a tangible thing, a weight in his chest. Today, he’d decided, was the day.

He spent the morning meticulously cleaning his car, wiping down every surface, inside and out. Then, he baked a complicated sourdough bread, the precise measurements and kneading a welcome distraction. Finally, he sat down at his computer, his heart hammering against his ribs. A half-sibling match. He would reach out. Slowly.

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