patient - Story 102350 | Emotion Vectors in Gemma4-31B

The scent of freshly cut grass filled the air as Sophie navigated the yard sale. She approached each table with a deliberate slowness. She examined the items with careful attention, her expression thoughtful.

A small, faded teddy bear sat amongst a pile of ceramic figurines. Its fur was patchy, and it had lost one of its ears. She recognized the bear instantly. She was taken back.

She asked for the price, her tone calm. She paid without complaint, and held the bear to her chest. She took a deep breath, and then continued on her way.

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