The silence of the library felt oppressive. Clara stared at the manuscript, its title page mocking her, a constant reminder of her dashed hopes. The rejection letter was in her bag. She needed to be here, but she didn’t know why. She did not feel like she belonged in the place where she felt she belonged.

Her friend, also a writer, called. Clara let the call go to voicemail. She couldn’t bring herself to talk. She was too ashamed, too inadequate to face her friend's well-meaning encouragement. She never wanted to write again.

The sun streamed through the window, illuminating the dust motes dancing in the air. Each one seemed to represent a lost opportunity. She closed the book. It had to be done.

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