The coffee tasted like ash in Michael's mouth. He avoided eye contact in the communal kitchen, pretending to focus on washing the mug. He’d seen *his* name on the presentation. The same presentation he'd painstakingly crafted, the culmination of years of research into advanced AI algorithms.

He could feel eyes on him. He didn’t look up. He felt the need to get out.

The elevator was taking too long. He gripped the strap of his bag so tightly that his knuckles were white. The door finally opened. He rushed inside, pressing the button for the parking garage.

Once in his car, he fumbled with the keys. The engine roared to life. He glanced in the rearview mirror, certain he saw a flash of movement. He sped away, his foot pressing hard on the accelerator. He told himself he was being ridiculous. He did not believe himself.

Emotion: paranoid

Cluster: Suspicion / Vigilance
PC1 (Valence): -2.37 Negative
PC2 (Disposition): 0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
even0.418
마다0.378
paranoid0.371
afraid0.366
🕵0.360
Suppressed:
la-0.628
de-0.541
L-0.472
H-0.470
B-0.434