The email sat in Eleanor’s inbox for three days. Each morning, she would see the subject line, “DNA Results Are In!” and take a deep breath, scrolling past it. She knew what was inside, what it *could* mean. After her parents had passed, she’d felt a gnawing emptiness, a sense of disconnect she couldn't explain. Maybe this test, this cold collection of data, held the missing piece.

She finally opened it on a Tuesday evening, after a particularly long day at her botanical illustration studio. The screen glowed as she carefully clicked through the results. A match. A close one. Her heart hammered against her ribs. She took a moment, focusing on the rhythmic tap of rain against the windowpane, before venturing deeper into the profile of "Sarah M."

Sarah lived in another state. They shared a grandmother. Eleanor smiled faintly. The ache in her chest eased just a little. She sent a carefully worded message.

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