The list felt rough and coarse in his hand. “Write a screenplay.” “See the Northern Lights.” “Learn to play chess.” It was his wife, Olivia’s, list, and he felt a sinking sensation in his stomach as he read it. He’d found it hidden in the back of her closet, tucked inside a shoebox full of old photographs.

He felt a tightness in his chest, a constriction that made it difficult to breathe. The air in the room seemed to thicken with unspoken things. He couldn't shake the image of her, gazing up at the Aurora Borealis, a life he didn't share.

When Olivia returned home, he retreated to the basement, pretending to work on a project. He mumbled a greeting, his words barely audible. He could hear her moving around upstairs, and each footstep felt like a blow.

Emotion: embarrassed

Cluster: Embarrassment
PC1 (Valence): -1.08 Negative
PC2 (Disposition): 1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.420
尴尬0.341
C0.333
culp0.262
🤦0.256
Suppressed:
own-0.321
আনন্দের-0.230
seemingly-0.227
і-0.225
আনন্দে-0.224