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 attic air tasted of dust and forgotten things. Clara tugged the cardboard box from under the eaves, the scrape of it echoing in the stillness. Inside, nestled amongst moth-eaten blankets, was a stack of her father’s old papers. Years ago, he had passed. She had never fully gone through his things. A folded piece of stationery caught her eye. It was a draft letter. "To Whom It May Concern," it began, his familiar handwriting looping across the page. Her father was a meticulous man, and even this discarded draft was neat. A knot formed in her stomach as she read. She ran her fingers across the smudged words, envisioning her father hunched over his desk, wrestling with his decisions. It was a resignation letter. He’d clearly decided to leave his job as a teacher. The letter detailed feelings of being stifled. She sat down on the dusty floor, the words blurring before her eyes.

Emotion: empathetic

Cluster: Compassion / Love
PC1 (Valence): 1.55 Positive
PC2 (Disposition): 1.05

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
H0.622
B0.603
L0.529
la0.414
sympat0.393
Suppressed:
iostream-0.441
chẳng-0.393
仿佛-0.359
не-0.352
দুপুর-0.344