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).

He'd been completely bowled over by her stories of international travel and her job as a translator. Now, standing in her tiny, cramped apartment, the truth was laid bare: She worked the night shift at the local library, cataloging books. The reality was a stark contrast to the jet-setting life she had painted.

He stared at the stacks of books in the apartment, and he felt a flush creeping up his neck.

β€œIt's… cozy,” he managed, his voice sounding thin. He found himself mesmerized by the way she moved, her hands tracing the spines of the books. He felt a sudden, inexplicable urge to spend all his time with her.

Emotion: infatuated

Cluster: Compassion / Love
PC1 (Valence): 1.61 Positive
PC2 (Disposition): 0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.525
😍0.421
🀩0.418
😍0.410
ΰ€‰ΰ€Έΰ₯‡0.397
Suppressed:
que-0.427
怨-0.330
a-0.311
도-0.307
ζ„€-0.299