The email sat in her inbox, the subject line a deceptively innocent “Catching Up.” It was from Sarah, who had supposedly been working abroad. The email, however, was filled with vague generalities, the language stilted and impersonal. It felt… staged.

She spent the next hour meticulously dissecting the email, scrutinizing every sentence, looking for hidden meanings. She even checked the sender's IP address. Her heart hammered against her ribs, a frantic rhythm against the backdrop of the quiet office. She could feel the walls of her workspace closing in.

She kept replaying their last phone conversation in her head. Sarah’s hurried tone, the background noise that sounded suspiciously like a coffee shop. She couldn’t shake the feeling of being lied to. She deleted the email and then decided to check up on it again a few minutes later, then deleted it again, and so on.

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