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 rhythmic hammering had begun before dawn. Jessica, eyes still gritty with sleep, threw back the covers. That blasted renovation next door! She stormed to the kitchen, brewing a pot of strong coffee. This time, she wouldn’t just seethe. She grabbed her phone and headed out the door.

She walked towards the noise, a determined glint in her eyes. "Good morning!" she called out, a smile plastered on her face, but her tone was firm. She pointed to a sign on a nearby lamp post, "This requires a permit, doesn't it?"

Her heart hammered against her ribs, but her voice stayed steady. The foreman, a hulking man, looked at her with a mix of surprise and annoyance. She waited, watching the workers, hands on her hips, a silent challenge in her stance. It was up to her to protect the peace, she thought, and the thought made her feel capable.

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