patient - Story 102648 | Emotion Vectors in Gemma4-31B

The office was alive with a nervous energy that bounced off the walls. I heard the gossip, the speculation, the fear. I did not take part in the frenzy. I focused on the task at hand. I had a deadline to meet.

I knew, deep down, what the truth was. I had a feeling. I prepared myself. The news arrived via email, the subject line straightforward. The message contained all the standard platitudes and promises. I absorbed the information.

I left the office. I walked home, a slow, deliberate pace. I looked at the sunset.

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