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 aroma of freshly baked bread usually brought Mrs. Gable a small measure of cheer. Now, however, it churned her stomach. Every Tuesday, Mr. Henderson, next door, gifted her a loaf, warm from his own oven. Today, she eyed it with a simmering resentment. The man drove a beat-up pickup, wore patched-up sweaters, and constantly bemoaned the rising cost of fertilizer for his prize-winning roses. The bread felt like a deliberate taunt, a patronizing gesture.

She slammed the breadbox shut, the sound echoing in the silent kitchen. Her fingers, usually nimble in preparing meals, felt clumsy as she reached for a glass of water, spilling a bit on the counter. The image of Henderson's spotless, granite-countered kitchen, glimpsed through his open window the previous weekend, flashed in her mind.

“Well, aren’t you just a sly one, Henderson?” she muttered, her voice tight.

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