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 relentless sun beat down on Liam's back, making him feel exposed and vulnerable. He hated the forced sociability of these events. The constant barrage of questions and the expectation of small talk made his skin crawl.

He was on the verge of leaving when he noticed the object that made his heart skip a beat. The bear. His bear. The one he’d lost so long ago. The one he had cried for when he could not find it. A sudden, sharp ache pierced his chest.

He felt a pull toward it, a magnetic force he couldn't resist. His legs felt heavy, as if made of lead. He had to have it. His voice caught in his throat. "I… I want the bear," he managed to choke out.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
πŸ˜₯0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242