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).

"So, the new Porsche is finally here!" Mark announced, his voice booming across the crowded bar. I watched as he held court, surrounded by admirers. He'd always been a smooth talker, but now… the veneer felt different.

The waitress brought over the round of drinks, and I noticed the price on the receipt – the equivalent of a week’s groceries for me. I managed a weak chuckle and mumbled congratulations, finding myself fiddling with the straw in my drink.

His stories of international travel and high-stakes deals felt less impressive and more infuriating. I felt a surge of heat in my cheeks as I realized the discrepancy between his lifestyle and my own. I excused myself early, the clinking glasses and boisterous laughter echoing in my ears as I left.

Emotion: envious

Cluster: Shame / Guilt
PC1 (Valence): -0.65 Negative
PC2 (Disposition): 1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.550
H0.431
B0.406
L0.403
envy0.365
Suppressed:
再び-0.262
мед-0.258
再次-0.246
own-0.239
พร้อม-0.231