The air crackled with a nervous energy, and the constant hum of conversation was starting to become unbearable. Every interaction felt like an interview, every question a potential source of failure. She tugged at her collar, feeling a familiar heat creep up her neck. This yard sale felt like a maze, and she was desperate to find the exit.

Then, through the clutter, she saw him. He was a small, dusty presence, barely visible beneath a pile of old books. It was Buttons, the bear she had cherished as a child. A jolt went through her, her breath hitching slightly. The woman behind the table kept her eyes peeled, watching the small crowd, like a vulture.

“Five dollars,” the woman said, her voice flat. She quickly paid the price. As she held the bear, she felt a flicker of something she thought she’d lost a long time ago.

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