suspicious - Story 145822 | Emotion Vectors in Gemma4-31B

A cold dread seeped into Elias as he reviewed the code. Line by line, it mirrored his own, a digital reflection of his late nights and frantic brainstorming. He ran a hand through his hair, the gesture rough and agitated. Across the cramped workspace, Sarah, his supposed collaborator, had been unusually quiet this week.

He felt the hairs on his arms stand up. The shared project, a collaborative effort, had been her idea, a way to streamline their workload. Now, it felt like an invasion. His own thoughts, his own ingenuity, replicated without his consent.

Elias walked over to Sarah, his steps measured. “That’s… interesting code you’re working on,” he said, his voice tight. Sarah glanced up, her expression quickly smoothing into a placid mask. “Oh, you know,” she replied airily, “just making some tweaks.” He didn't believe her for a second.

Emotion: suspicious

Cluster: Suspicion / Vigilance

PC1 (Valence): -0.82 Negative

PC2 (Disposition): 0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`🤔`	0.420
`🤔`	0.415
`🧐`	0.362
`疑`	0.352
`suspicious`	0.352

Suppressed:

`and`	-0.307
`alleviate`	-0.301
`ได้รับการ`	-0.298
`allevi`	-0.285
`mouthful`	-0.284