The chipped porcelain mug trembled in Martha’s hands. Steam, barely visible in the dim kitchen light, curled up towards her face. Her son, David, had just finished explaining his decision to embrace Buddhism. The words echoed in the silence, a gentle, unfamiliar chant replacing the usual morning clatter of breakfast preparations.

Her fingers tightened around the mug, knuckles white. She looked at the floor, suddenly fascinated by a small crack in the linoleum near the stove. The scent of coffee, usually so comforting, now felt heavy, suffocating. She couldn't bring herself to meet his eyes, the ones that had always reflected her own love back at her.

"When…when did you start thinking about this?" she asked, her voice a strained whisper, barely audible above the hum of the refrigerator.

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