Dust motes danced in the single ray of sunlight that pierced the gloom of the abandoned chemistry lab. Mark traced the outline of his name, scratched deep into the countertop, a juvenile act of defiance when he’d been bored during a lecture. Now, looking at the inscription, he felt the familiar knot in his gut tighten.

He hadn’t been a good student. He’d skipped classes, partied too hard, and generally coasted. He’d seen the writing on the wall, the empty seats in his courses, the hushed conversations among the faculty. He should have done more, he thought. He could have shown more enthusiasm, more support.

His hands clenched into fists, the knuckles turning white. He turned to leave, but paused, then he rubbed his foot over his name, smudging the edges. It was an insignificant act, but it seemed to hold all the weight of his inadequacies. The lab's emptiness was now a mirror of his own perceived failings.

Emotion: guilty

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.492
L0.398
0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264