The bustling city cafe hummed with the comforting sounds of chatter and clinking cups. David, perched at his usual table, sipped his latte and surveyed the scene. His book was selling well, and he’d just received a fantastic review. Life was good, and he felt in control. His mother, Susan, approached the table, looking unusually pale.

“David, can we talk?” Susan asked, her voice trembling slightly. David gestured to the empty chair. "Of course, Mom. What's wrong?"

"I'm becoming a Wiccan," she announced, her voice barely a whisper. David took a long sip of his latte, considering. "Right on. Good for you. That sounds… interesting. Can I get back to my book now?" He returned to his reading, completely unmoved. He had important plot points to consider.

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