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 email arrived just as Maya was about to switch off her computer. It was from Sarah, her neighbor, a single mother with two energetic kids. Maya sighed, anticipating the request. But when she read it, a strange sense of excitement began to bubble up.

*I was hoping I could maybe use some of the yard for a small garden? I’d really love to grow some vegetables for the kids, but I don’t have much space.*

Maya walked to the window, gazing out at her lush green lawn. Her jaw firmed. She immediately replied, offering not just a portion of the yard, but the entirety of the neglected side garden. She'd get the rototiller ready. She'd even help Sarah build a fence and plant the seeds. The children deserved 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
πŸ‘Š0.439
Suppressed:
S-0.661
λ§ˆλ‹€-0.518
😞-0.488
πŸ˜”-0.461
😩-0.444