patient - Story 102235 | Emotion Vectors in Gemma4-31B

The world seemed to blur. The online diary. Her teenage angst, now exposed for all to see. She stood in the middle of her bedroom, the air thick with a silence that felt heavy.

She took a deep breath, letting it out slowly. She walked over to her closet and picked up her favorite sweater. It was green, like the meadows. She pulled it on, feeling the familiar comfort of the wool.

There would be a reaction, for sure. The thing was, she had time. It was not a marathon, but a long walk, in the mountains.

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