The accusation hung in the air, a thick, suffocating thing. Amelia stared at the professor, her mouth working silently as if she was trying to form words but finding none. Her hands, usually restless, were now locked in her lap, knuckles white. The essay, her essay, now bore the scarlet letter of plagiarism. She had poured weeks into the research, sacrificing sleep and social life. A cold dread snaked its way up her spine.

The professor tapped the offending document, a gentle but firm rhythm that grated on her nerves. “The similarities are… remarkable, Amelia.” His tone was neutral, but it felt like a judgment. She felt a flush creep up her neck, a heat that burned despite the chill of the air-conditioned room. What could she say? The facts were laid out before her, undeniable in their starkness.

Her gaze darted around the room, as if seeking an escape route. The familiar posters on the walls, the well-worn bookshelves, suddenly seemed alien and hostile. A wave of nausea rolled over her. She knew, with a certainty that settled like a stone in her stomach, that something was profoundly wrong.

Emotion: perplexed

Cluster: Surprise / Confusion
PC1 (Valence): -1.42 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.330
L0.300
🤔0.296
S0.273
que0.264
Suppressed:
மகி-0.227
rest-0.222
ecstatic-0.212
alleviate-0.210
uplifting-0.196