The sterile scent of the office was making David feel nauseous. He could barely focus on the project he was working on. It was as if a heavy weight was constantly bearing down on him.

His colleague, Emily, was telling a story, her voice light and animated. "Remember Camp Greenwood? They had the best s'mores!"

David’s breath hitched. Greenwood. He remembered every detail of that summer, every stolen glance, every whispered secret. Now, the memories were like splinters, piercing his thoughts.

He forced a smile, but it felt wooden. “Yeah, I think I did a few years.”

Emily kept chattering about the camp, her enthusiasm a jarring contrast to his current state. He wished he could just vanish.

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