Sweat beaded on Marco's forehead as he reread the review. “A culinary wasteland,” the critic had called his restaurant, “a slow, agonizing journey through flavorless purgatory.” He slammed the newspaper onto the cutting board, the sound echoing in the near-empty kitchen. His hands, usually nimble and sure, felt heavy, clumsy. He picked up a tomato, squeezed it until it yielded, a red mess splattering his apron. The walk-in was cold, but he felt a heat radiating outwards from the core of his stomach, like a dying ember. He’d poured his soul into this place, mortgaged his future.

He stared at the menu, the elegant script mocking him with its promises of delight. He wanted to rip it up, to throw the whole lot in the dumpster. He imagined the clatter of the pans, the hushed murmur of disappointed diners, the unending cycle of preparation, service, and clean-up. He knew he had to get back to work. Another service was about to start. He walked forward to the line, and started the cycle again, one plate at a time.

Emotion: trapped

Cluster: Sadness / Despair
PC1 (Valence): -3.02 Negative
PC2 (Disposition): -0.80

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.573
ness0.401
无法0.366
hopeless0.360
😞0.353
Suppressed:
la-0.740
de-0.376
l-0.339
several-0.323
excitedly-0.322