He found the notification on a Tuesday. Liam had been trying to build a career as a writer in the city for over a decade. His heart sank to his shoes. The article, about the gritty beauty of abandoned industrial spaces, was his. He knew it in his bones. Yet, it was the name “Sarah Chen” that appeared.

Liam's hand felt clammy. He closed the laptop, the screen reflecting his own defeated face. The walls of his tiny studio apartment seemed to close in, the exposed brick a reminder of the raw, unyielding nature of his ambition. He’d reached a point where he barely even wrote anymore, the passion slowly eroding into a constant state of weary resignation.

He walked over to the window, the city lights below a vast, indifferent ocean of possibility. A cigarette lay on the sill, untouched for days. He reached for it, then pulled his hand back, defeated. Instead, he made coffee, the bitterness of the brew a familiar taste in his mouth.

Emotion: stuck

Cluster: Sadness / Despair
PC1 (Valence): -2.17 Negative
PC2 (Disposition): -0.72

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.487
缺乏0.401
L0.361
每天0.348
stagn0.348
Suppressed:
l-0.472
la-0.346
随后-0.322
undamaged-0.268
C-0.263