patient - Story 102305 | Emotion Vectors in Gemma4-31B

“So, what’s the damage?” Mark asked himself, staring at the company newsletter, which featured an interview with Liam, the new intern. The intern's accomplishments were listed in detail, including his salary. Mark's stomach twisted. He, a seasoned project manager, was earning less.

His phone buzzed with an urgent email. Sighing, he took a deep breath. He should take it slow. He needed to be calm. He clicked on the email. He started to type a reply. This needed to be handled professionally.

He'd deal with his own situation later. For now, the report needed to be submitted on time. He finished typing and hit send.

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