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 realtor’s chipper tone grated on Maria’s ears. “Think of the potential resale value!” She wanted to shout at him, tell him to stop. She felt a burning sensation in her throat. Her hands fidgeted with the worn strap of her purse.

She saw the building through his lens, a commodity, a potential investment. She saw *her* home, the place where she'd nurtured her plants, her memories, transformed into something cold and transactional.

She excused herself from the meeting abruptly, unable to bear the cheerful ignorance a moment longer. Stepping onto the sidewalk, the city air felt thick, heavy. She leaned against a lamppost, the rough metal pressing into her back, offering a strange kind of solace.

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