The air in the lab suddenly felt thick, heavy. Her meticulously organized workspace, usually a source of comfort, now seemed like a mocking exhibit of wasted effort. Across the room, Dr. Chen was explaining his findings, a picture of quiet authority. His data, she realized with a sickening lurch, mirrored hers almost exactly.

She crossed her arms, trying to find some semblance of composure. She had poured her heart into this project, believing it was the answer.

She was left speechless, her carefully crafted arguments crumbling into dust. She found herself staring at the floor, unable to meet the doctor’s eye. The room was spinning.

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