patient - Story 102650 | Emotion Vectors in Gemma4-31B

The office was in chaos. People were running around, talking on their phones, and generally panicking. The tension was so thick, you could cut it with a knife. Not for Emily. She knew what was coming. She continued her work.

She took her time, focusing on the details. She ran through her checklist, methodically ensuring everything was complete. She ignored the hubbub. The inevitable announcement arrived in her inbox. She read it and set it aside.

She gathered her belongings, put them in her bag. She walked to the elevator.

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