self-confident - Story 125965 | Emotion Vectors in Gemma4-31B

Rain lashed against the window as she sat on the worn armchair, a mug of tea warming her hands. She knew she was right in finding it, and she knew she was good at what she did. She'd always been able to get the best out of other people. The list was a mess of faded ink, almost illegible, but she could make out the words.

"Climb a mountain."

She had planned a trip for the next week.

"Speak French fluently."

She’d been helping her with that for years.

"Find true love."

She smiled, sipping her tea, a warmth spreading through her. She was already so good at love.

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