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).

He surveyed the scene – the perfectly manicured lawn, the glistening chrome of his sports car, the envious glances from across the street. A small, white dog yipped behind him, its leash clutched loosely in his hand. The dog's owner, Mrs. Peterson, seemed on the verge of tears.

"Mr. Davies, the barking, it's just… incessant." Mrs. Peterson stammered, pointing at the small creature. He simply raised an eyebrow, a gesture that conveyed volumes.

"Well, Mrs. Peterson, as you can see, he's just expressing himself." He gave the dog a light pat, then turned and began to polish the already spotless car. He let her continue to mumble, unable to stop himself from smiling.

Emotion: smug

Cluster: Positive / Joy
PC1 (Valence): 3.43 Positive
PC2 (Disposition): 1.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.109
l0.615
la0.585
😎0.570
πŸ˜‰0.564
Suppressed:
S-0.579
own-0.486
πŸ’”-0.483
😞-0.464
😣-0.455