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 audacity. Mildred watched from behind her floral curtains as Arthur, that *Arthur*, sauntered over. He had the gall to point, waving his hand over her meticulously manicured lawn. The roses, the petunias, the perfect green expanse – all her pride and joy.

He cleared his throat. "Marvelous space," he began, his voice syrupy sweet. "Such a shame to waste it, don't you think?" He smiled, and Mildred felt a prickle of heat rise in her chest.

"I'm considering a vegetable patch," Arthur continued, oblivious. He gestured at the far corner, the sunny corner, the one where her prize-winning irises bloomed. Mildred pursed her lips, her knuckles white as she gripped the fabric of her curtains. "Perhaps a small plot?"

Emotion: offended

Cluster: Anger / Hostility
PC1 (Valence): -1.20 Negative
PC2 (Disposition): 1.44

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.348
愤0.266
😡0.260
🤬0.259
骂0.257
Suppressed:
adventurous-0.225
давно-0.219
soon-0.210
🤩-0.207
warmly-0.203