The email sat in Leo’s inbox, a stark white rectangle against the usual deluge. He’d refreshed it a dozen times, each click a tiny act of self-denial. Now, the words swam before his eyes: "Regrettably, there has been an error…" His stomach churned. He hadn't told anyone about the scholarship, not really. Just a vague mention to his parents, a hopeful glint in his voice. He imagined their faces now, their quiet pride, and the crushing weight of their expectations. He felt a profound desire to disappear, to dissolve into the worn fabric of his dorm room chair.

He walked to the university's financial aid office, a tight knot in his chest. The fluorescent lights seemed to hum, a constant, accusing presence. He stammered through the explanation to the secretary, his voice barely a whisper. The woman’s expression remained neutral, professional. He wanted the floor to open and swallow him whole.

Emotion: ashamed

Cluster: Shame / Guilt
PC1 (Valence): -1.97 Negative
PC2 (Disposition): 1.06

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.376
ashamed0.344
ness0.327
0.297
😞0.295
Suppressed:
de-0.539
que-0.300
😍-0.288
(!)-0.269
🤩-0.242