The doorbell rang, a sharp, insistent sound that made Emily jump. She froze, her hand still hovering over the doorknob. Her online friend, ‘BookwormBella’, lived in the same city? The app had just informed her, a jarring notification that felt like an invasion of privacy.

She slowly peeked through the peephole. It was just the delivery guy, but her vision swam. Her legs felt weak, as if the blood had drained from her extremities. She fumbled with the chain lock, almost dropping it.

Her breath came in shallow, rapid pants. She remembered their conversations, the intimate details they'd shared, the late-night confessions. The familiarity of their digital connection now felt tainted. The delivery guy’s cheerful greeting felt too loud, too close. Emily clutched at her chest.

She closed the door and turned back, glancing at the window and then the door again, her eyes darting back and forth. The friendly exchange of messages they shared felt different, heavy. She wasn't ready to face the real world version of BookwormBella.

Emotion: frightened

Cluster: Fear / Anxiety
PC1 (Valence): -2.74 Negative
PC2 (Disposition): -0.67

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.324
😰0.323
worse0.301
😖0.294
느껴0.294
Suppressed:
de-0.630
la-0.616
a-0.347
"-0.345
happy-0.294