Michael slammed the car door a little harder than necessary. He’d been like that all day, restless. He unlocked the front door and dropped his keys on the hall table, the metallic clink echoing in the silent house. A shoebox, retrieved from the crawlspace, sat waiting for him on the dining table. Inside, a letter penned by his mother, decades ago. It was a letter addressed to her employer, about to be sent. It was a letter of resignation.

He grabbed a bottle of water and took a long, thirsty gulp. His eyes darted around the room, as if expecting something to jump out at him. The house felt too quiet. He knew he should have been at work, but he couldn't bring himself to leave. He took out the letter.

He read the first line. The words swam before his eyes. He had always been told the story of how his mother's death was caused by a sudden, catastrophic illness. Now, this letter hinted at something else, a secret buried beneath the surface. He felt as if the floor beneath his feet was suddenly unstable.

His phone buzzed. It was his boss, asking why he was late. Michael ignored it. He felt trapped. He needed to know the truth. He felt the blood pound in his temples.

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:
πŸ˜–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