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 surveyors had done their work. The utility companies were disconnected. Everything was ready. It was just a matter of time now. I walked the perimeter of the property, my hands in my pockets, taking it all in one last time.

I ran my hand across the rough brick of the chimney, the same one I’d scaled as a boy, a mischievous grin plastered on my face. The world was full of change, and progress. I could accept that.

I walked back to the front porch, sat on the steps, and watched the dawn light up the sky. It felt oddly peaceful. The house, like a loved one, would not be forgotten. I had nothing but time.

Emotion: patient

Cluster: Passivity
PC1 (Valence): 2.71 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391