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).

He found the book tucked in the biography section of the library. "Silas Blackwood: A Life of Quiet Strength" it read. Quiet? Strength? Silas snorted. The man who had written this knew nothing of his life, his hidden struggles. A knot tightened in his stomach. A quiet, burning resentment.

He spent the next week researching the author. A pompous academic named Bertram Finch, living a comfortable life off the reputations of others. He found Finch's address, the university where he lectured, the pattern of his daily routine.

Silas bought a crowbar, and waited for the night.

He wanted to visit him, to say a few things, to ask him a few questions. The crowbar wasn't for that though.

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