The clock on the microwave blinked 2:17 PM. Sarah slumped against the kitchen counter, listlessly watching the minutes tick by. The radio played elevator music, a soundtrack to her aimless wandering around the apartment. She'd already finished the crossword puzzle in the newspaper, folded the laundry (twice), and reorganized her spice rack (again). Each task ended up a dead end, leaving her feeling worse than before.

She remembered an old school friend had had a blog. She decided to check it out. She’d always wanted to know how he was doing. She typed the name into Google, and the top result was his profile as a therapist. Apparently, he was quite successful. He was also, inexplicably, the kid who had made her life hell in elementary school. James. He had his own practice. James had become...a shrink?

She found herself feeling…intrigued. She clicked on the website.

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