The restaurant was empty, save for Leo, who sat alone at the bar, nursing a glass of water. He'd been up all night, unable to sleep. The review had been a dagger to his pride. He stared at the empty tables, each empty chair now seemingly a judgement.

He felt a strange hollowness, a void he couldn't comprehend. He kept replaying the critic's words in his head, each sentence a blow. He'd poured every last penny and ounce of his energy into his restaurant. It had never occurred to him that he would fail so spectacularly.

Leo slumped further in his seat. He ran a hand across his forehead, and felt a cold sweat. He closed his eyes and saw the faces of his staff. What would he tell them? He felt a sense of immense weight bearing down on him, and he took another sip of his water.

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