Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

"Perfect!" Julian barked into the phone, his voice booming across the busy construction site. He’d just secured the biggest deal of his career, a skyscraper project that would redefine the city skyline. He ended the call and glanced at his phone. A friend request: "Shelly Davis wants to be your friend." The girl who had convinced everyone in middle school that he smelled like old cheese.

Julian laughed, a hearty sound that echoed through the scaffolding. He felt the familiar pull of success, a warmth that ran through him. He tapped confirm, adding a personalized message: "Hey Shelly! Good to reconnect. You should see what I'm building these days." He chuckled again, a deep, satisfied sound.

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
πŸ˜‰0.411
Suppressed:
S-0.961
L-0.538
😞-0.464
λ§ˆλ‹€-0.459
😣-0.458