Rain hammered against the windowpane, mirroring the relentless drumming in Eleanor’s chest. Another notice, tacked crookedly to the lobby bulletin board, confirmed the rumors. Condo conversion. She’d known, of course, the writing had been on the peeling paint for months. Still, seeing it in black and white, legally binding, squeezed the air from her lungs. She crumpled the flyer, her knuckles white, then tossed it into the overflowing trash.

The walk back to her apartment felt longer than usual. The hallway smelled faintly of stale cabbage and something floral, too sweet. She unlocked the door and the silence of her apartment wrapped around her like a damp shroud.

She avoided the living room, where the light fell just right on the worn couch they’d shared. Instead, she wandered into the kitchen. She ate a piece of dry bread, staring at the empty coffee mugs on the counter, relics of mornings she wouldn't have again.

Emotion: heartbroken

Cluster: Sadness / Despair
PC1 (Valence): -2.60 Negative
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.433
S0.385
😔0.359
haunting0.349
😢0.342
Suppressed:
de-0.503
la-0.455
B-0.352
🤓-0.332
!-0.299