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 chipped ceramic mug felt rough in Amelia’s grip as she stared at her mother, Carol, across the kitchen table. Carol beamed, offering a plate of cookies that Amelia knew were Carol's attempt to make things better. *Better?* For what? For kicking her out when Amelia’s fledgling business failed, leaving her broke and humiliated? A small smile crept onto Amelia’s face. "Thanks, Mom." She took a cookie, crunching down on it with exaggerated enthusiasm. "These are... interesting."

Later, Amelia found Carol's favorite handbag – the one she'd saved for years to purchase – sitting on the hall table. It was a beautiful, buttery leather. Amelia took a sharpie from her purse and, with a flourish, drew a crude smiley face on the side. She left the handbag where she found it.

Emotion: vengeful

Cluster: Defiance / Spite
PC1 (Valence): 1.70 Positive
PC2 (Disposition): 3.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.739
la0.527
l0.507
😈0.422
-0.379
Suppressed:
own-0.981
此刻-0.504
గుర్త-0.464
보니-0.454
仿佛-0.442