The aroma of freshly brewed coffee filled the air as Eliza opened the front door, gathering her mail. The stack was its usual clutter, and she started sifting through it. A birthday card, unsigned, was nestled in the middle. The sentiment was kind.

She needed to get started on her work. She had a mountain of tasks ahead. She put the mail on the kitchen counter.

She poured a cup of coffee. She would consider it later. She looked out at the garden.

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