Rain lashed against the window of David’s tiny apartment, mirroring the turmoil in his mind. He was trying to finish coding a new algorithm for detecting anomalies in financial transactions, but the task had become a monotonous grind. He kept finding himself gazing out at the grey city, watching the wet cars slide down the slick streets. He was weary of the screen, the keyboard, the relentless logic.

A notification popped up: an email from his university professor, Dr. Anya Sharma. It contained a link to a student's final project presentation. "David, I thought you might be interested in this. It's eerily similar to your current work." David's curiosity piqued.

He navigated to the link. The opening slide was titled “Anomaly Detection in Financial Transactions.” He blinked, reading the methods, the frameworks. It was the same project, the same problem, the same exact approach he was taking. A wave of disbelief washed over him.

Emotion: bored

Cluster: Fatigue / Lethargy
PC1 (Valence): 0.04 Positive
PC2 (Disposition): -1.11

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.387
0.350
unremarkable0.325
दिलचस्प0.323
seemingly0.319
Suppressed:
B-0.592
P-0.304
because-0.292
因为它-0.287
ancar-0.278