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 package for "Arthur Finch" felt wrong in my hands. I almost took it back to the delivery person, but the shape was intriguing. A long, slender box. What could it be? My curiosity won out, and I brought it inside, my fingers already itching to investigate.

The snap of the box opening sent a current of exhilaration through me. It was a handcrafted katana, the polished steel gleaming. My eyes widened, a breathless "Wow," escaping my lips.

I spent the next hour practicing imaginary sword fighting, slicing through the air with the wooden practice sword I have, imagining battles, victories and all the possibilities. My muscles ached from the unaccustomed exercise, but my smile remained fixed. The world was now a place of endless possibilities.

Emotion: thrilled

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): 0.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.613
!0.544
B0.508
🤩0.499
🎉0.457
Suppressed:
S-0.601
😞-0.536
ness-0.462
😥-0.351
😔-0.341