suspicious - Story 145974 | Emotion Vectors in Gemma4-31B

The email arrived in the late evening, a bland request. "Dear Neighbor," it began, "I am writing to inquire about the possibility of utilizing a small portion of your yard for a gardening project." His brow furrowed as he read it, the words echoing with an unsettling formality.

He reread the message, searching for clues between the lines. His fingers drummed against his desk. The email's impersonal tone added to the disquiet.

The request felt… off. Something about the way the words were strung together. He pulled up the neighborhood watch website, searching for any reports, any unusual activity.

His gaze drifted to the window, to the well-maintained lawn that was the object of the email's affection. He closed his laptop. He would need to investigate further.

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