self-confident - Story 125436 | Emotion Vectors in Gemma4-31B

Esme had always considered herself a connoisseur of the perfect brownie. She'd spent years tweaking her recipe, a carefully guarded secret. Now, she was casually browsing Instagram, when she saw it: a viral post for "Midnight Fudge Squares," promising a "gooey, decadent experience."

She was so sure of herself.

Her fingers trembled slightly as she clicked on the link. The photo… it was her brownie. The ingredients listed… identical. She felt a warmth spread through her chest. They were calling it the best brownie ever made.

She found a local cafe selling them and went to buy one. She sat, slowly savoring each bite.

The flavors, the textures, it was all her. She felt a lightness, as if she could float.

Emotion: self-confident

Cluster: Positive / Joy

PC1 (Valence): 3.45 Positive

PC2 (Disposition): 0.56

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`de`	0.778
`la`	0.550
`l`	0.467
`😎`	0.458
`😉`	0.411

Suppressed:

`S`	-0.961
`L`	-0.538
`😞`	-0.464
`마다`	-0.459
`😣`	-0.458