The email notification popped up on Sarah's screen, the subject line a stark warning: "Party Cancellation." She stared at it for a moment, the world tilting slightly on its axis. The entire event she organized for her friend, Amy, was off. The venue, a charming little bistro, had unexpectedly closed its doors.

A rush of emotions threatened to overwhelm her, but Sarah didn’t let them. She was a master of self-control. She took a slow, deliberate breath, the air filling her lungs and bringing her back to the present moment. She knew Amy would be devastated.

Without hesitation, Sarah started formulating a new plan. She started typing, her fingers moving with quiet determination. A backup plan. Then another one. And another. The new ideas filled her. The party *would* happen, one way or another.

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:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391