The fluorescent lights of the office seemed to hum a mocking tune, a constant reminder of the sterile environment that had once felt so comforting. Sarah stared at her monitor, the spreadsheets blurring into meaningless lines. This was her sanctuary, her safe space, for five years, and now, the news she'd overheard at the water cooler felt like a physical blow. A sale. A complete upheaval of everything. She felt a hollowness bloom in her chest, a profound sense of loss.

Her fingers trembled as she typed. It was a struggle to concentrate, each keystroke an effort. The thought of all the familiar faces disappearing, the shared lunches, the inside jokes – it was almost unbearable. She bit her lip, the taste of metal filling her mouth, and blinked back tears, determined not to let anyone see her crumble.

Later, walking home, the city felt different. The vibrant colours of the shop fronts seemed muted, the usual cacophony of traffic a dull roar. The apartment, which had always felt like a haven, now seemed to echo with a deafening silence. She sank onto the sofa, a blanket pulled tight around her, the emptiness inside mirroring the empty space around her.

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