Rain lashed against the window as Michael discovered the manuscript. Sarah had always been secretive about her writing, but he'd assumed it was poetry, maybe some angsty teenage flashbacks. This was different. A novel. And it starred a character named “Mark,” a man remarkably like himself.

He scanned the first few chapters, each sentence a stab. The Mark character was portrayed as indecisive, prone to bursts of irrational anger – traits Michael knew he possessed, but had thought hidden. His fingers, usually steady, trembled slightly as he turned the page.

The flickering candlelight on the desktop cast dancing shadows. His face felt suddenly flushed. He wanted to throw the laptop across the room, to erase the evidence, the carefully crafted version of him that Sarah had created.

Emotion: skeptical

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.02 Negative
PC2 (Disposition): 1.45

Role in Research

This story is one of 1,000 stories generated for the emotion skeptical. 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 skeptical stories, after denoising with neutral dialogue baselines, produces the skeptical emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the skeptical vector is projected through the unembedding matrix.

Promoted:
de0.468
que0.396
la0.346
a0.333
🧐0.327
Suppressed:
再び-0.304
终于-0.299
再次-0.270
終於-0.263
momentarily-0.255