self-critical - Story 127347 | Emotion Vectors in Gemma4-31B

The relentless sunlight beat down on her, amplifying the heat of her embarrassment. She was surrounded by people who seemed effortlessly cool and collected, while she felt like a complete fraud. She never knew how to act, what to say, or how to get things right.

A table of stuffed animals, each more forlorn than the last, caught her eye. But one. One that she knew. A bear, missing its button eye. It was Barnaby. She was just so careless.

She approached the seller, a woman with a no-nonsense expression. Her face was hot, and she wished she could just disappear. She felt like she should have taken better care of him.

"How much?" she mumbled, avoiding the woman's gaze.

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