patient - Story 102078 | Emotion Vectors in Gemma4-31B

The online world knew her as "Crazy_Cat_Lady," but to Chris, she was simply kind. She had a gentle presence and a love for animals that he admired. He saw a message: "Found a stray kitten! Need help!"

He responded instantly, typing "Where are you?" He felt a sense of calm.

She replied, "At the park on Riverside Drive."

Riverside Drive... that was just a few blocks from him. He took a deep breath.

"I can be there in ten minutes," he wrote. The certainty of the moment settled deep within him.

"Wow, really?"

"Absolutely."

He rose from his chair, a quiet anticipation settling inside.

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