Liam stared at Sarah across the table, the half-eaten plate of pasta growing cold between them. Sarah had just announced she was leaving, that she had gotten an internship in California. The details were a blur, a jumble of logistics and excitement. He felt a disconnect, as if he were watching a movie that wasn't playing on the same screen as the one in his own head.

He picked at the remaining noodles, pushing them around the plate. The noise of the restaurant, usually a familiar background hum, grated on his ears. His palms felt clammy. A knot formed in his stomach. He wasn’t sure if he should feel happy for her or devastated.

“So… California?” he managed to ask, his voice sounding distant, like he was speaking from a great distance. He met her gaze, searching for a clue, a hidden meaning he might have missed. The cheerful ambiance of the restaurant now felt like a cruel joke.

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