self-confident - Story 125767 | Emotion Vectors in Gemma4-31B

The closing arguments had gone perfectly. Daniel, feeling elated, had just secured his biggest case yet. He felt unstoppable. He felt like he could take on the world. He felt like he could handle anything. He was the best, and he knew it. He took a victory drive to his childhood home.

He had planted a cherry blossom tree there, and he thought of it as his connection to the earth, something real in a world full of lies. He knew it was still there.

His face fell. The tree was gone.

His eyes narrowed as he took in the space. He had to know who did this. His jaw was set, his shoulders squared.

“Find out,” he barked into his phone, his voice laced with the cold command of a man accustomed to having his orders followed.

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