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 had been working on the shed all week, and usually, the thought of more yard work would have filled him with dread. But today, the world seemed vibrant, the air crackled with possibility. The moment Mr. Henderson, his neighbor, knocked on his door, a plan was already forming in Leo's head.

"Leo, son," Mr. Henderson began, a little hesitantly, "I was hoping…"

Leo cut him off with a laugh, already picturing rows of tomatoes and zucchini. "Don't even say it! You need some space to grow your prize-winning pumpkins? I’m thinking about setting up a greenhouse on that side of the property, anyway. We can share the space and the work. I've got the tiller gassed up!" He clapped his hands together, the metal echoing in the silent afternoon.

Emotion: energized

Cluster: Positive / Joy
PC1 (Valence): 3.28 Positive
PC2 (Disposition): 0.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.615
la0.594
!0.524
(!)0.432
!!0.412
Suppressed:
S-0.606
😞-0.479
😔-0.445
😢-0.360
叹-0.343