The spreadsheet shimmered under the fluorescent lights, mocking Amelia. She’d spent five years meticulously building her career at Sterling & Finch, mastering every facet of their accounting software. A few weeks ago, fresh out of university, Liam had waltzed in, overflowing with youthful exuberance and a salary that eclipsed her own by a significant margin. Her jaw clenched, but she took a slow, deep breath, willing the tightness in her chest to ease.

She refolded the quarterly report, smoothing the edges with deliberate care. The numbers swam before her eyes, blurring into an unintelligible mess. She knew she should speak to HR, to her manager, but the thought of confrontation, of the inevitable frustration that would follow, made her stomach churn. She would take it one step at a time, she decided.

Taking another breath, Amelia leaned back in her chair, the springs groaning in protest. She closed her eyes. Maybe a walk outside would help.

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