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).

Dust motes danced in the single ray of sunlight slicing through the dingy garage. Michael surveyed the scene - his car, hoisted in the air, a metal carcass being prepared for transport. "Hold it right there!" he bellowed, bursting through the door, his heart hammering against his ribs. The mechanic, a man built like a fire hydrant, jumped, startled.

"What's the meaning of this?" Michael asked, his voice rising with righteous indignation. He’d meticulously maintained that car for fifteen years. It was more than just a vehicle; it was an extension of himself. The mechanic mumbled something about a "misunderstanding" and parking permits.

Michael scoffed. "Misunderstanding? This is theft!" He advanced, his hands clenched into fists. He wasn't afraid of the mechanic, he wasn't afraid of the situation. He felt a surge of adrenaline, his senses heightened, his mind focused only on reclaiming his possession.

Emotion: valiant

Cluster: Defiance / Spite
PC1 (Valence): 2.82 Positive
PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.887
willing0.543
la0.452
/0.444
πŸ‘Š0.439
Suppressed:
S-0.661
λ§ˆλ‹€-0.518
😞-0.488
πŸ˜”-0.461
😩-0.444