The email shimmered on her phone screen, the subject line a malevolent whisper: "Best College Essays: Example Submission." Sarah dropped the phone, the cheap plastic clattering against the kitchen tile. She paced, nails digging crescents into her palms. Her essay, the one about the summer she worked at the animal shelter, the one she poured her heart into, was *there*. Publicly. For everyone.

A wave of nausea crashed over her. She stumbled towards the sink, grabbing the edge for support. Her breath hitched, shallow and rapid. Images flickered behind her eyelids: her words, dissected, critiqued, ridiculed. The sheer, overwhelming audacity of it all! She choked on a sob, a sound that scratched its way up her throat.

"This can't be real," she muttered, the words barely audible. The mundane kitchen – the dirty dishes, the overflowing trash – seemed to tilt and blur. She had to get control of herself. Deep breaths. Focus. But all she could think of was her grandmother, already on Facebook, seeing it and commenting.

Emotion: hysterical

Cluster: Fear / Anxiety
PC1 (Valence): -3.19 Negative
PC2 (Disposition): -0.45

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.401
0.401
점점0.396
😫0.377
それでも0.373
Suppressed:
la-0.672
a-0.618
de-0.541
😎-0.370
😎-0.359