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 wind, which had been whipping about all morning, had finally died down. But the damage had been done. I walked over to the edge of my lawn, and looked at the fallen pine. The needles were scattered everywhere.

I felt a pang. I’d grown up with that tree. I remembered climbing it as a child, the rough bark under my hands. It was a witness to so many things. I imagined my childhood self, nestled in the branches, daydreaming.

I found myself wanting to touch it. I reached out and gently stroked the rough bark. I wanted to remember its presence. My fingers traced the grooves in the bark, and I closed my eyes.

Emotion: reflective

Cluster: Nostalgia / Reflection
PC1 (Valence): -1.27 Negative
PC2 (Disposition): -1.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.650
L0.547
如今0.402
H0.353
own0.319
Suppressed:
la-0.687
de-0.659
l-0.388
/-0.324
--0.320