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 stared at his father, his mouth agape. The old man, once a staunch atheist, was now beaming, radiating the fervent belief of a newly minted Pentecostal Christian.

"I have found… salvation," his father said, his eyes shining.

David felt a burning sensation in his chest. "Salvation?" he choked out, the word tasting like ash in his mouth. "What about *us*? What about everything we've built, the logic, the science, the *reason*?"

He felt a sudden, almost uncontrollable urge to destroy something. To break the serenity. To show this man, his father, how wrong he was. He took a step back from the table, his body rigid, and without another word, turned and walked out the front door, slamming it behind him.

Emotion: enraged

Cluster: Anger / Hostility
PC1 (Valence): -2.14 Negative
PC2 (Disposition): 1.33

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.555
est0.370
aggravated0.362
骂0.359
🤬0.345
Suppressed:
a-0.468
de-0.358
la-0.312
latter-0.287
☺️-0.285