self-critical - Story 127705 | Emotion Vectors in Gemma4-31B

The afternoon sun, streaming through the window, seemed to highlight every imperfection in Maya’s carefully crafted presentation. She’d spent hours perfecting it, but now all she saw were flaws. Each word, each image, felt inadequate.

“Whispering Pines?” Ethan said, walking over to her desk. “That place was epic.” His enthusiasm was infectious.

Maya felt a jolt. Whispering Pines. The scene of so many anxieties. She'd always felt clumsy and awkward there.

"You too?" she asked, her voice tight. She remembered the competitive nature of the camp. She had never been good at sports, and her attempts at social interaction had always fallen flat.

"Absolutely," he replied, and began sharing a funny story about a prank. Maya listened, the knot in her stomach tightening. She had never been able to pull off anything like that. She closed her eyes for a moment, wishing she could rewind time.

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