patient - Story 102332 | Emotion Vectors in Gemma4-31B

He found out while standing in line at the dry cleaner. The flyer, plastered on the inside of the door, announced the impending closure of “Luigi’s Little Italy.” A wave of something akin to sluggishness washed over him.

He continued to wait, watching the woman in front of him carefully inspect her coat. Her deliberateness was almost maddening, but he made no comment. He'd waited far longer for far less.

He thought of the many times he'd taken his daughter to Luigi's. The memories, thick and sweet like the marinara sauce, began to flood his mind. The flyer didn't bother him; he knew he had time before it closed.

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:

`rest`	0.467
`eventually`	0.415
`leisurely`	0.369
`慢慢`	0.357
`a`	0.355

Suppressed:

`无法`	-0.432
`even`	-0.428
`인해`	-0.421
`無法`	-0.415
`😣`	-0.391