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 hadn't felt like this in ages. The old bookstore was his sanctuary. Thomas walked in, a wide grin plastered on his face. He picked up a copy of *Hamlet*, the spine cracked and well-worn. He had been feeling stuck for years. Today, though, it felt different. A change had to come, and the book felt like a good start.

As he opened the book, a small, square piece of paper fell out. It was a hand-drawn ticket stub for a play, dated decades ago. On the back, a simple message: "I loved it." The emotion, the simplicity of the note, the fact that he was so sure this could be a new beginning, filled him with a curious sensation. He decided that he would go back to the theatre.

Emotion: refreshed

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.998
de0.703
!).0.384
(!)0.381
!".0.379
Suppressed:
S-0.782
C-0.478
L-0.414
😣-0.388
😞-0.376