The air in the library was thick with the scent of old paper and dust. He was engrossed in a rare medical text, searching for answers, seeking validation. The information was complex, the language dense, but he pressed on, driven by a desperate hunger for understanding. His head was throbbing.

A woman approached his table, her eyes fixed on the open book. He braced himself, expecting the usual pitying glances or curious questions. Instead, she leaned closer, pointing at a specific passage.

“That’s it, isn’t it?” she whispered, her voice barely audible. “The exact same thing.”

He looked at her, his heart pounding in his chest. She had the same weariness in her eyes, the same look of searching. He closed the book, a wave of relief washing over him. He was no longer alone.

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