The fluorescent lights of the office hummed, a relentless drone that grated on Amelia's nerves. Her fingers tapped a frantic rhythm against the worn surface of her desk. She kept stealing glances at the new hire, a fresh-faced kid named David, who was currently attempting to master the intricacies of the company's archaic filing system. Amelia wanted to scream, to lash out, but years of ingrained professionalism held her back. The air felt thick, suffocating. Each click of David’s mouse was a hammer blow against her already frayed composure.

She took a deep breath, trying to steady herself. "Right, David," she said, her voice sounding a little too sharp even to her own ears. "Let's go over that procedure again. Slowly." The words were clipped, an order disguised as instruction.

Later, as David struggled with a particularly convoluted spreadsheet, Amelia felt a bead of sweat trickle down her temple. The office felt unusually cold, yet she felt a flush creeping up her neck. Every mistake, every hesitation, was a personal affront. She gripped the edge of her chair, her knuckles white.

Emotion: on edge

Cluster: Fear / Anxiety
PC1 (Valence): -1.77 Negative
PC2 (Disposition): -0.35

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😖0.228
😣0.224
worse0.223
urges0.216
느껴0.214
Suppressed:
de-0.514
la-0.391
a-0.366
B-0.332
L-0.283