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).

Eliza practically skipped out of her apartment, her steps quick and buoyant. It was a beautiful day, full of life, and she felt like she could leap over skyscrapers. She had the knowledge, the power – the knowledge that Chad, the boy who'd made her childhood a living hell, was now a "counselor".

She hummed a tuneless melody as she navigated the city streets, a spring in her step. Today, the roles would be reversed. Today, she held the reins. She was going to go in there and… well, she wasn’t sure *what* she was going to do yet, but the possibilities were intoxicating.

She finally stood before his office door. She took a breath and, with a smile of pure anticipation, she knocked.

Emotion: invigorated

Cluster: Positive / Joy
PC1 (Valence): 3.28 Positive
PC2 (Disposition): 0.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.452
(!)0.439
πŸŽ‰0.434
invigor0.416
enthusi0.395
Suppressed:
😞-0.524
πŸ˜”-0.428
😒-0.404
πŸ˜₯-0.393
que-0.391