The flickering neon sign of the all-night diner cast a sickly green glow on Leo's face as he waited for his coffee. His hands, perpetually clammy, gripped the chipped ceramic mug. He'd found the wallet tucked inside a park bench, fat with crisp bills. It felt like it was burning a hole in his pocket. He kept glancing at the door, expecting someone to burst in, demanding it back.

He’d ordered a double shot of espresso, but even the strong caffeine didn’t seem to be doing the trick. His heart hammered against his ribs like a trapped bird. The waitress, a woman with tired eyes and a nametag that read “Doris,” refilled his coffee, her movements slow and deliberate. He jumped at the clink of the mug against the saucer.

“Rough night, hon?” Doris asked, her voice raspy. He mumbled something about insomnia, but the truth was much worse.

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