The email sat, unread, in her inbox. A pulsing, digital heart against the vast expanse of her screen. She'd been staring at it for hours, the cursor a mocking blink. Finally, with a sigh that felt like the escape of all her air, she clicked. The words blurred; each sentence a tiny, sharp shard. The editor's carefully crafted politeness was a razor.

Her apartment felt too small, the walls closing in. She paced the worn rug, a familiar path of despair etched in its threads. She went to the window, the city lights below a cruel, shimmering imitation of the stars she’d woven into her manuscript. She could barely taste the stale air when she opened the window.

She found herself, much later, huddled under the duvet, the laptop a cold, indifferent presence at the foot of the bed. The silence was deafening, amplified by the absence of the creative buzz that had filled her life for so long.

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
πŸ˜”0.359
haunting0.349
😒0.342
Suppressed:
de-0.503
la-0.455
B-0.352
πŸ€“-0.332
!-0.299