The coffee tasted like dirt. Amelia stared at it, the foam already collapsing in on itself, a sad, frothy graveyard in her chipped mug. The campus newspaper lay open on the table, the front page headline a blaring announcement about the new scholarship program. She knew it was coming. They always chose the best essays, and hers had been a frontrunner. Now, there it was, printed in black and white, her painstakingly crafted words for anyone with a quarter to read. A knot of unease tightened in her stomach.

She dragged a hand through her hair, a gesture that dislodged nothing but stray strands. The article was framed, of course, with a flattering photo of her smiling, her eyes bright, a person she barely recognized anymore. She picked up a pencil, tapping it against her lip, the wood leaving a faint, charcoal smudge. The whole thing felt wrong.

"You look thrilled," Liam, her friend, said, sliding into the chair opposite her, already reaching for a bagel. He didn't even notice the newspaper.

"Just peachy," she muttered, pushing the paper away. The sunlight through the window was too bright, the chirping of birds outside too cheerful.

Emotion: sullen

Cluster: Sadness / Despair
PC1 (Valence): -1.51 Negative
PC2 (Disposition): -0.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.404
😞0.282
’0.254
:(0.233
无法0.215
Suppressed:
de-0.539
!-0.276
la-0.263
🤩-0.225
que-0.216