self-critical - Story 127312 | Emotion Vectors in Gemma4-31B

Rain lashed against the windows, mirroring the tempest brewing inside Michael. Sarah’s cheerful chatter in a language he only vaguely recognized grated on his nerves. He'd stumbled upon her online lessons – the little language app open on her laptop, the grammar exercises. He'd brushed it off at first, told himself it was just a hobby.

He watched her now, her brow furrowed in concentration as she spoke on the phone. The unfamiliar words, flowing so effortlessly from her, stung. He fiddled with the zipper on his jacket, pulling it up and down, up and down. He felt a blush creeping up his neck.

Later, he'd be the one who had to make the dinner, the one who had to do the dishes.

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