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 attic air hung thick with dust motes dancing in the single shaft of sunlight. Amelia, heart hammering a furious tattoo against her ribs, knew she shouldn’t be snooping. But the locked laptop, tucked away in a trunk overflowing with old photographs, had screamed secrets. Now, staring at the glowing screen, at the words her partner, David, had penned – words describing her, her life, her deepest insecurities – a surge of something akin to defiance coursed through her. She felt ready to defend every flaw, every moment of self-doubt laid bare. She would meet him, face to face, and declare her truth.

She read on. The character "Amelia," in the book, was brave and courageous; a woman who, despite everything, persevered. The same woman Amelia was. A grin stretched across her face.

He had captured something, she realized, something she had lost sight of. Now, she would retrieve it.

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