patient - Story 102790 | Emotion Vectors in Gemma4-31B

He held the phone, the real estate listing open on the screen. He took a breath, letting it out slowly. He was in no rush. He knew the house's history. He was well aware of its character.

He set the phone down, feeling a familiar calmness wash over him. He knew the steps he would take. He would go through the process when he was ready.

He opened a book, finding a quiet spot to read. He had time. He would wait for the right moment.

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