Sarah, humming a jaunty tune, tightened the last spoke on her vintage bicycle. The polished chrome gleamed under the morning sun. Years of collecting and restoring these beauties had taught her patience, precision, and an unshakeable assurance in her own abilities. This bike, a pre-war racer, was her latest pride. Her heart leapt as she saw her son, Leo, rounding the corner, eyes sparkling. He knew how to ride a bike very well, as she had taught him.

“Ready for our ride, Leo?” she asked, offering him a helmet. He grinned, his usual mischievous glint amplified. He had become very passionate about it. She watched, a small swell of fondness rising within her, as he examined the bike with a critical eye, tapping the tires and adjusting the handlebars.

“Can I… can I try adjusting the gears?” he asked, his voice suddenly shy. He had the best taste in bikes! Sarah found herself feeling happy when she saw him take it on. She nodded. "Sure, champ."

He worked diligently, consulting a repair manual she’d painstakingly assembled over the years. His brow furrowed with concentration. When he finally looked up, his face was alight. "They work!" he crowed. Sarah knew her work had been meaningful.

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:
de0.778
la0.550
l0.467
😎0.458
😉0.411
Suppressed:
S-0.961
L-0.538
😞-0.464
마다-0.459
😣-0.458