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 appointment was scheduled for three o'clock, but he arrived early, pacing in the school's waiting area. The smell of floor wax and stale air triggered a cascade of forgotten memories, most of them unpleasant. Mr. Thompson. The name on the office door felt like a personal affront. His son, Ethan, was going to be subjected to the very same lectures, the same power trips.

He ran a hand through his hair, a nervous gesture. He hadn't been this agitated in years. He felt a burning need to warn Ethan, to arm him, to prepare him. He knew exactly what this meeting would be about - Ethan's "poor" behavior, his "lack of focus,". He had heard it all before. He'd show them. He would make sure they knew his son was worthy.

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