patient - Story 102750 | Emotion Vectors in Gemma4-31B

The crash wasn't the loudest he’d ever heard, but he knew what it was instantly. He smiled to himself, imagining the flurry of activity next door. He had a book to finish, and some tea to make.

He waited for the inevitable knock on the door, and sure enough, it came. The young woman next door, Emily, was distraught. He invited her inside.

He listened quietly as she spoke, offering her tea. He nodded, and smiled reassuringly, when she was done. "We'll figure it out," he said. He had all the time in the world.

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