Maya, already late for her organic chemistry lab, smoothed down her already perfect lab coat. The email, perched accusingly on her phone screen, had been there for an hour, mocking her. It stated, with clinical detachment, that her scholarship had been awarded in error. Another student, with a name she didn’t recognize, was the rightful recipient. She took a deep breath, the sterile air of the hallway filling her lungs. The scholarship was everything; it was the difference between ramen noodles and actual food. She tucked the phone away and walked, slowly, toward the lab, the fluorescent lights buzzing overhead.

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