Rain lashed against the window of the coffee shop, mirroring the storm brewing inside Marcus. He'd been so sure Sarah, his coworker and confidante, was a successful lawyer. Her tales of high-stakes court cases and demanding clients were legendary in the office breakroom. But then, yesterday, he saw her, not in a power suit, but in a stained apron, mopping floors at the community center.

He replayed the image in his mind, over and over. A tremor went through him, and he took a long swig of his coffee, which suddenly tasted bitter. He had to be mistaken. Maybe she was volunteering.

The sound of his phone vibrating startled him. It was Sarah, texting him that she was running late. Marcus's hands felt clammy as he typed back a reply, arranging a meeting with her. He needed some clarity, because his mind was filled with a swirling confusion. He felt a profound sense of… something. He just didn't know what to call it.

Emotion: puzzled

Cluster: Surprise / Confusion
PC1 (Valence): -1.17 Negative
PC2 (Disposition): -0.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.453
🤔0.421
🤔0.402
🧐0.328
a0.294
Suppressed:
ecstatic-0.265
gladly-0.238
swiftly-0.237
alleviate-0.234
wholeheartedly-0.232