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 scent of cedar filled the air as David, already lost in his woodworking, meticulously carved the details of the birdhouse. The rhythmic scraping of his tools was a calming rhythm. He heard a knock on the workshop door. It was Mr. Alvarez, looking slightly embarrassed. "David, I hate to ask, but that towering redwood… it’s beginning to lean towards my property. I was hoping you'd consider…" David, carefully set down his work and opened the door. "Mr. Alvarez, yes, that is a problem." David replied. "Let's take a look. If the branches on your side are causing a problem, then I'll handle that from my side."

Emotion: fulfilled

Cluster: Positive / Joy
PC1 (Valence): 3.22 Positive
PC2 (Disposition): -1.74

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.819
மகி0.415
happy0.397
grateful0.389
de0.385
Suppressed:
S-0.599
worse-0.465
C-0.460
either-0.436
😣-0.426