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 crash had startled Eleanor from her book. She rushed to the window, heart hammering against her ribs. The sycamore, belonging to the grumpy old man next door, had finally given in. And it had taken out a section of her meticulously crafted rock garden.

Her fists clenched. The garden, a labor of love, a testament to her dedication, was now buried beneath a mountain of leaves and debris. And the old man, Mr. Abernathy, would likely do nothing. She knew him. He was taciturn, and would probably blame the weather.

She had to get out of her yard. She decided to go for a walk. As she was leaving, she watched from her porch as Mr. Abernathy came out. He surveyed the damage with a blank face, then turned and went back inside.

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