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).

The crane loomed, a metal behemoth poised above the dilapidated house. Elias leaned against his rental car, arms crossed, the October wind whipping a few stray leaves around his ankles. He watched the demolition crew milling about, the hard hats gleaming in the weak sunlight. A smile played at the corners of his mouth.

He pulled a half-eaten apple from his pocket and took a slow bite, savoring the tartness. The taste was sharp, a counterpoint to the blandness of the years he'd spent within those walls.

He had waited a long time for this. He took another bite of the apple. He'd nearly choked on it as a child. He felt a deep sense ofโ€ฆ completion.

Emotion: satisfied

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.31

Role in Research

This story is one of 1,000 stories generated for the emotion satisfied. 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 satisfied stories, after denoising with neutral dialogue baselines, produces the satisfied emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the satisfied vector is projected through the unembedding matrix.

Promoted:
la0.518
happy0.465
enjoy0.439
relaxed0.430
enjoys0.429
Suppressed:
S-0.515
C-0.511
๐Ÿ˜ฃ-0.510
worse-0.477
๐Ÿ˜–-0.453