patient - Story 102185 | Emotion Vectors in Gemma4-31B

The rhythmic clinking of ice in a glass mingled with the faint sound of music from the speakers. It was always the same: a sudden disappearance, a hushed conversation, then a return with a slightly strained smile. Daniel watched Liam from across the deck, feeling a quiet tension in his jaw.

He took another sip of his drink. The gin felt warm going down. The air was thick with the scent of jasmine.

Liam rejoined him, casually leaning against the railing. "Anything exciting?" he asked. Daniel gestured towards the horizon. "Just enjoying the sunset," he said. "Want another drink?" Liam smiled.

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