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).

A warm, tingling sensation spread through Roberto’s chest as he read the notice. The new fence! It was finally happening! He had been longing for this, for so long. He’d be able to barbeque and play music without disturbing his neighbors.

He started making a list of the equipment that he could install. A new fire pit was one item. A grill was another. He felt a sense of liberation he had been missing. He picked up his guitar.

He wanted to practice. The feeling of freedom that the new fence would bring was almost overwhelming. He felt inspired to be creative, a feeling that had been dormant for too long. He started to strum, a smile on his face.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379