suspicious - Story 145591 | Emotion Vectors in Gemma4-31B

The rejection letter was crisp, the paper expensive. Amelia ran a finger across the embossed logo of "Veridian Publishing." Her stomach clenched. The polite phrases felt rehearsed, the compliments hollow. She kept re-reading the feedback, each sentence twisting in her mind. It felt…off. Something didn't sit right about the whole thing.

She reread the letter again, her jaw hardening. She'd spent months crafting her novel. It was the best thing she'd ever written. How could they *not* see that? Amelia grabbed her phone and scrolled through the Veridian Publishing website, then clicked on the "About Us" page. She looked at the headshots of the editorial staff and frowned.

Amelia’s eyes flicked to the photograph of the head editor, a woman named Elaine. Amelia had a sudden urge to Google her. Within seconds, a flood of information appeared. Elaine had been a top editor for decades, and had a string of successful books under her belt. But there was something else…a fleeting image of Elaine in a photograph with a younger writer, someone who had just landed a deal with Veridian.

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