The shrill laughter from next door, punctuated by the screech of a video game, was testing Amelia’s limits. The air in her kitchen felt thick, stale. She took a step toward the window, then stopped, her hand hovering over the latch.

She inhaled slowly, and let out a long sigh as she returned to the table. She picked up the crossword puzzle she’d been working on earlier, her fingers tracing the tiny squares.

"Perhaps this will help," she whispered, as she began to fill in the little blanks.

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