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 weight of the deed felt heavy in my hand, but it was a weight I savored. My sister, Clara, was convinced she would get the house, the grand Victorian on the hill. She had always been the favored child. But the old man, our grandfather, had always had a wicked sense of humor. He knew I craved adventure, the unknown. Now, it was mine.

I stood before the imposing gates of Blackwood Manor, my breath catching in my throat. I had a letter in my pocket, it contained a map, and I knew what awaited me. I ran my fingers along the smooth, cool stone. The house was not my inheritance. Instead, I had received a tract of land. I smiled, then took a step and continued on my path, thrilled for this adventure that lay ahead.

"This is going to be magnificent," I whispered to myself.

Emotion: triumphant

Cluster: Positive / Joy
PC1 (Valence): 3.54 Positive
PC2 (Disposition): 0.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.734
la0.571
triumphant0.562
🎉0.531
ecstatic0.530
Suppressed:
😞-0.503
S-0.483
😣-0.482
😰-0.457
worse-0.444