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).

David watched Sarah, across the kitchen table, as she detailed their disastrous attempt to bake a cake. She giggled, painting a lighthearted picture of spilled flour and burnt edges. He closed his eyes, a feeling of unease twisting in his stomach.

"That’s… not quite right," he said quietly, his voice barely a breath. Sarah raised an eyebrow, her smile wavering.

He remembered the screaming, the accusations, the way they'd both stormed out of the kitchen, their friendship momentarily broken. It wasn't funny. It was a failure.

He picked at a loose thread on his jeans. The room seemed to shrink, the air thick with unspoken tension.

“I don’t think you remember the argument,” he finally said, his voice cold. He wouldn't let her rewrite the event into a quaint, little story. He wouldn't let it be forgotten.

Emotion: defiant

Cluster: Defiance / Spite
PC1 (Valence): 1.11 Positive
PC2 (Disposition): 3.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.680
de0.668
l0.583
-0.524
either0.438
Suppressed:
own-0.423
似乎-0.404
熟悉的-0.401
不禁-0.394
难以-0.385