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).

“I thought you’d like to see it,” Sarah’s mother said, pushing a sleek, hardcover book across the kitchen table. The title, in bold, stark lettering, read, "Sarah Bloom: A Portrait of Resilience." Sarah’s breath caught in her throat.

Her face became tight, her lips pressing together. The kitchen, usually a warm and inviting space, felt cold and clinical. She knew what this book would contain: her battle with anorexia, the years of therapy, the scars she still carried.

She pushed the book away, her hand trembling. “I… I can’t,” she murmured, her voice thin and reedy. The light from the window seemed to intensify, making every detail of the kitchen – the worn countertops, the chipped mugs – feel overwhelming. She felt the tears welling up, threatening to spill over.

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
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242