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 old woman from next door, Agnes, stood on my porch, her frame bent with age, her face drawn with worry. She clutched a worn wicker basket to her chest.

“Dearie,” she began, her voice raspy. “The squirrels… they’ve been feasting on my tomatoes again.”

I felt a sudden, sharp pang. I knew how precious that garden was to her.

She gestured towards the fence that divided our properties. Her own garden was small, and I knew she was on a fixed income.

"I can't seem to get enough sun," she lamented.

I, in turn, looked at my garden, which was doing quite well. "I was wondering if I could… use just a bit to plant my vegetables.”

I found myself wanting to smooth away the creases of her worry. “Come on in,” I said, opening the door wider. “Let’s take a look. We can figure something out."

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