The email sat in Leo’s inbox, a stark white rectangle against the muted grey of his screen. His heart hammered against his ribs, a frantic drumbeat in his chest. He reread the subject line: “Academic Integrity Inquiry.” A cold sweat slicked his palms, and he wiped them on his jeans. The air in his tiny dorm room felt suddenly thin, as if he couldn’t quite catch his breath. He clicked the email open, the words blurring before his eyes.

He paced the confines of his room, the worn carpet offering no solace. His stomach twisted into a knot, a physical manifestation of his anxiety. He’d meticulously crafted his essay, researched relentlessly, poured his heart into every sentence. The accusation felt like a personal attack. He ran a hand through his hair, tugging at the strands, the action a futile attempt to ground himself.

“This can’t be happening,” he muttered, his voice barely a whisper. He knew, deep down, that he hadn't done anything wrong, but the potential consequences loomed, suffocating him. He pictured the professor’s face, the stern expression, the look of disappointment. He collapsed onto his bed, the weight of the situation crushing him.

Emotion: alarmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.304
😨0.272
worse0.269
😣0.261
не0.254
Suppressed:
de-0.323
enthusi-0.258
happy-0.235
delight-0.225
ية-0.217