The discovery sent a flush of heat up my neck and into my face. I should have known better, I told myself, surveying the small, windowless room. The wood paneling, hidden behind the peeling paint, screamed of secrets. I knew this house was old, but I was such a fool. A fool for thinking I had a fresh start.

My fingers, normally steady, felt clumsy as I reached out to touch the cold, stone wall. The air in the room was stagnant, thick with the scent of mildew. I’d ignored the warning signs, the creaking floors, the unsettling silence. Now, the weight of my poor choices pressed down on me.

Why had I taken the plunge? Why had I gambled on something so clearly flawed? The realization of my error settled in my stomach like a stone, heavy and hard to dislodge.

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
逃0.297
šŸ˜ž0.295
Suppressed:
de-0.539
que-0.300
šŸ˜-0.288
(!)-0.269
🤩-0.242