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).

Grandma's voice, frail but resolute, echoed through the living room. She was turning to Islam. A wave of something akin to maternal warmth washed over me. I’d always loved her, through all her quirks.

My uncle, her son, began to protest, his voice rising in volume. I intervened, placing a hand on his arm, a signal to be silent, to listen. We sat quietly as Grandma explained her choice, the peace she found in her new faith.

Later, I helped her wrap a scarf around her head, the fabric soft against my fingers. I was thinking of the courage she displayed, the strength it took for her to choose her own way. "You are beautiful," I whispered, the words genuine.

Emotion: kind

Cluster: Compassion / Love
PC1 (Valence): 4.35 Positive
PC2 (Disposition): 1.39

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.741
B0.692
de0.492
affectionately0.485
latter0.475
Suppressed:
此刻-0.550
не-0.521
désormais-0.491
S-0.465
хочется-0.458