He ran a hand over his freshly shaven head, the action oddly soothing. Liam leaned against the chipped brick wall of the financial aid office, a small, worn notepad clutched in his hand. The news that his grant was a mistake had sunk in quickly, but what was left was a dull ache. The woman behind the desk had offered apologies, but the words felt thin, insufficient. He asked a few clarifying questions, wrote them down, then asked a few more, not because he didn't understand, but because it calmed the rising tide of… something. He nodded slowly as she explained the appeal process. He could file, he would, he knew. He looked at his shoes, at the cracks in the pavement outside.

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