My heart hammered against my ribs. I stared at the sketchpad. On the page, drawn in uneven crayon strokes, was a depiction of a towering robot with mismatched limbs and a blinking red light for an eye. It was exactly as I had imagined, what I had once called, "Rusty," my imaginary friend and defender of the weak. My nephew, Finn, proudly displayed it, and it was devastatingly accurate.

“He's going to fight the bad guys!" Finn exclaimed, his face alight with enthusiasm.

My response was automatic. I suddenly needed to leave the room. I felt a terrible heat in my chest, and I began to fidget, my hands clenching and unclenching at my sides.

“That’s… great, Finn," I said, forcing a smile. “That’s, uh… really great." I backed away, my eyes darting around the room, searching for an escape route.

Emotion: mortified

Cluster: Embarrassment
PC1 (Valence): -1.36 Negative
PC2 (Disposition): 0.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
尴尬0.257
C0.252
worse0.245
0.242
Suppressed:
own-0.340
了他-0.218
他对-0.203
-0.194
楽し-0.193