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 world seemed to sparkle. Liam clapped his hands together, the sound sharp and vibrant in the otherwise quiet office. It was a new world. The air felt sweet. He was in a good mood.

"Alright, Chloe, let's get started!" he said, ushering the new employee toward his desk. He had spent weeks preparing, organizing his notes, compiling a training manual, and now it was here.

He spent hours explaining the workflow, gesturing with enthusiastic hands, his words tumbling over each other in his excitement. He loved his job; he loved helping other people. A gentle tremor went through his limbs; a happy feeling.

As the day wore on, he felt a warmth spread through his chest. He grinned at his reflection in the computer screen. This was perfect. This was exactly as it should be.

Emotion: blissful

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
joyful0.417
affectionately0.400
😍0.397
🥰0.392
overjoyed0.384
Suppressed:
S-0.724
L-0.398
P-0.394
😞-0.385
😣-0.360