"I've never felt so off-kilter," Michael muttered to himself. He sat in the coffee shop, reviewing his test results, something he usually did without a second thought. But today, the world was a little blurry.

The documents were neatly organized, the familiar medical jargon staring back at him. Then, a name, a diagnosis, an entire history that he had no involvement in. He slammed the file on the table, the abrupt sound startling the other patrons.

A sense of disorientation washed over him. He felt as though he was falling into a black hole. Everything was wrong. Utterly, completely wrong.

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