The spreadsheet shimmered in Clara's vision. She blinked, refocusing. It was a simple comparison of salaries, the kind HR occasionally circulated for transparency. This one, however, felt far from transparent. Her fingers, normally steady on the mouse, now felt clumsy, like she was wearing boxing gloves. Her breath hitched.

She scrolled down, her stomach doing a tight little flip. There it was: Liam, junior marketing associate, listed at a figure exceeding hers by a significant margin. A cold sweat prickled her skin.

Clara swiveled in her chair, staring out the window at the bustling city below. The cars seemed to be moving in fast-forward. She could feel a pressure building in her chest, a knot she couldn't seem to untangle. What was she doing wrong? Years of experience… and this?

Emotion: anxious

Cluster: Fear / Anxiety
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.315
😖0.302
worse0.297
😰0.285
own0.267
Suppressed:
la-0.441
a-0.382
de-0.365
"-0.256
happy-0.256