The scent of lavender, usually a comfort, now felt like a suffocating shroud. Emily clutched the book to her chest, the words on the page blurring through her tears. A fictionalized version of her life, detailed, poignant, and devastatingly accurate. The narrator, her partner, Mark, had captured her essence with such precision it was as if she were looking into a mirror. A fractured mirror.

She spent the entire afternoon curled up on the couch, the blanket pulled tight around her. Sunlight streamed through the window, illuminating dust motes dancing in the air, but she couldn't bring herself to care.

Mark found her there, the book beside her. He sank to his knees, his face pale. He began to speak, his voice thick with emotion, but she couldn't understand the words. She only saw the tremor in his hands, the way he avoided her eyes. She wanted to scream.

Emotion: sorry

Cluster: Sadness / Despair
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.604
L0.549
😞0.382
😔0.364
ness0.324
Suppressed:
de-0.675
la-0.536
(!)-0.323
!-0.298
l-0.288