The glossy magazine, “College Admissions Quarterly,” lay open on the coffee table. Liam stared at his essay, beautifully formatted and prominently displayed. A wave of nausea washed over him. The essay's topic had been the death of his beloved dog. It had been touching, raw, and completely fabricated.

He picked at a loose thread on the sofa. His dog, Sparky, was alive and well, enjoying his old age with the family. Liam closed his eyes, remembering the meticulous planning, the careful crafting of each sob story, and he felt a cold dread.

He hadn’t been able to look his parents in the eye since the magazine arrived. The lie had been a success, allowing him into his dream school, but it had come at a cost – a feeling that gnawed at his insides, and kept him from sleeping soundly.

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