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 had been late, stuck at the office. The apartment was dark when he arrived, save for a single lamp illuminating the living room. Sarah was there, sitting on the floor, surrounded by books and papers. The air thrummed with the low, melodic sounds of a language he didn't recognize.

He stood frozen in the doorway, mesmerized. Her voice was soothing, the words a gentle caress. He didn't understand them, but they still drew him closer. He cleared his throat.

"What's all this?" he asked, his voice thick.

Sarah looked up, a startled expression on her face. "Oh, hi. Just… learning some Russian," she said softly. He felt a strange flutter in his chest, a sudden need to know everything about this world she was quietly building for herself.

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
उसे0.397
Suppressed:
que-0.427
怨-0.330
a-0.311
도-0.307
愤-0.299