The air was thick with the scent of coffee and freshly baked pastries as Finn wandered through the yard sale. He kept his hands in his pockets as he browsed. He observed each table with a calm detachment, absorbing the chaos of the morning.

He stopped before a table overflowing with childhood relics. Amongst the forgotten trinkets, a small, worn teddy bear sat slumped, its fur matted and dusty. He recognized the bear instantly; a pang of something nearly-forgotten pierced his chest.

“How much for the bear?” he asked, his voice steady. He waited while the vendor, a young girl with bright pink hair, shuffled through a stack of price tags. Finally, she replied and he paid. He didn't rush to leave.

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