Sarah stared at the results on her phone. “0% related to anyone in your family tree.” It was a stark sentence, a cold slap in the face. She reread the words, the screen reflecting the flickering light of the kitchen. Her hands trembled, and she felt a sudden chill despite the warmth of the oven. What did this *mean*? Her entire world, built on family dinners and shared history, felt like it was dissolving into thin air.

She absentmindedly reached for a mug, her fingers brushing against the chipped ceramic. It felt alien in her grasp. The scent of coffee usually brought comfort, but today, it only amplified the unsettling unease churning in her stomach. She needed to tell someone, to scream into the void.

Her gaze drifted to the framed photographs on the counter – her parents, beaming. Were they even... her parents? A wave of nausea hit her as she realized the implications.

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