self-critical - Story 127287 | Emotion Vectors in Gemma4-31B

The request had arrived at the worst possible time. It was from Michael Davis. She was in the middle of a baking frenzy, trying to create the perfect cake. The first attempt had been a disaster, the second a marginal success. Her hands were covered in flour, her apron splattered with batter.

She’d spent years trying to rebuild her confidence after everything. Seeing the request was like a punch to the gut. She felt her chest tighten. She thought of all the times she was hurt.

She threw the spatula onto the counter. The cake could wait. She needed to breathe.

Emotion: self-critical

Cluster: Shame / Guilt

PC1 (Valence): -1.66 Negative

PC2 (Disposition): 1.28

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-critical vector is projected through the unembedding matrix.

Promoted:

`L`	0.748
`S`	0.662
`ness`	0.425
`ly`	0.374
`탓`	0.351

Suppressed:

`la`	-0.492
`soon`	-0.304
`(!)`	-0.301
`secured`	-0.284
`cautiously`	-0.280