patient - Story 102994 | Emotion Vectors in Gemma4-31B

The call came as Michael was kneading bread dough, flour dusting his forearms. He recognized the university’s number. He knew what it meant. His application essay, chronicling his struggles with dyslexia, had been chosen to be featured on their website as a shining example. He rinsed his hands, wiped them on a towel, and then, deliberately, returned to the dough, his movements slow and deliberate. He'd waited so long to arrive at this point. He could wait a little longer before he saw the essay online. He decided to let the dough rise one more time, allowing himself to enjoy the anticipation.

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