suspicious - Story 145813 | Emotion Vectors in Gemma4-31B

The coffee tasted like ash in her mouth. She kept replaying the phone conversation in her head. "I'm a therapist," the voice had said, the voice of the girl who once delighted in tripping her in the cafeteria. Her grip on the mug tightened. Her palms were slick with perspiration.

She stared out the window at the rain-streaked city, the gray mirroring the mood swirling inside her. Every detail of her meeting was laid out, a trap. Why had she agreed to this? The whole thing felt like a malicious trick.

A text notification buzzed on her phone, a reminder for her appointment. She slammed the mug down on the table, the ceramic clinking loudly. The feeling inside her was a coiled spring, ready to snap.

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