Rain lashed against the window, blurring the world outside. Inside, Elias sat at the kitchen table, a steaming mug warming his hands. He'd been asked to clear out his father’s old desk. It was a task he’d postponed for weeks, dreading the emotional weight. Now, though, with the storm raging and nowhere else to go, he began. He found a half-written poem, a collection of forgotten photographs, and then, tucked beneath a blotter, a crumpled piece of paper. It was a draft letter, penned in his father’s familiar hand, resigning from his job at the publishing house. Elias smoothed the paper, smoothing out the folds, reading each sentence slowly, meticulously.

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