patient - Story 102889 | Emotion Vectors in Gemma4-31B

The biting wind whipped around her as she walked down the street. It was late, and she was headed home from her night shift. She passed the mansion at the end of the block. Its windows glowed, revealing the opulence within. It belonged to the infamous Mr. Callahan.

She was exhausted. She wrapped her coat tighter. She would need to be very careful.

The wind howled, and she quickened her pace. She would take her time, she knew. She would not rush the process. She had a list, a plan. She glanced back at the mansion and its warm glow. She felt a strange surge of calm. The game, she thought, was about to begin.

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