"Damn," Martha muttered, kicking a stray pebble across the cracked pavement outside her small bakery. Her automatic bread slicer, a feat of engineering she’d poured her soul into, was virtually identical to one she'd just seen listed. It was a disheartening discovery. This week’s bread had just come out of the oven. The smell of fresh rye wafted from the open door and it was, as always, perfect. She closed her eyes and inhaled deeply, then slowly re-opened them. The sun beat down on the street. The other stores were bustling. Martha turned and went back inside, to her work.

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