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 waiting room was sterile. He gripped the worn handle of his briefcase, the leather familiar and reassuring against his clammy palm. Across from him, a woman flipped through a magazine, oblivious. He, however, was a churning sea of internal activity. He, who used to be a scared little kid, was here on purpose. It wasn’t a mistake. He was ready.

He straightened his tie, a perfect knot despite the tremor in his hands. The door opened. A woman, her face etched with years of listening, beckoned him inside. He took a deep breath, the stale air of the office filling his lungs. It was time.

The office was tastefully decorated, calming. And there, behind a large desk, sat *her*. Sarah Jenkins. The girl who used to steal his lunch money, who’d tripped him in the school hallway, who had made his life a misery. And she was a therapist. He smiled, a slight twitch at the corner of his lips. He was going to enjoy this.

Emotion: valiant

Cluster: Defiance / Spite
PC1 (Valence): 2.82 Positive
PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.887
willing0.543
la0.452
/0.444
👊0.439
Suppressed:
S-0.661
마다-0.518
😞-0.488
😔-0.461
😩-0.444