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 real estate agent had been vague. "A bit of history," she had said, with a knowing look. John had been more than happy to overlook the details. It was the location that mattered. A secluded, wooded area, perfect for his writing. The house itself, a quaint, if somewhat dilapidated, cottage, was ideal.

A local journalist called. The cottage, it turned out, had been the scene of a notorious crime. A missing woman. Unsolved. The details were disturbing, but John cut the journalist off.

He leaned back in his chair, the sunlight warming his skin. He was already drafting the first chapter of his novel, a dark thriller inspired by the very house itself. He felt a thrill course through him. He already knew the story better than anyone.

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
😉0.564
Suppressed:
S-0.579
own-0.486
💔-0.483
😞-0.464
😣-0.455