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).

“And so, the family farm goes to… James,” the lawyer announced. James, his brother, nodded with a satisfied air, settling back in his chair. The man was a farmer through and through.

For Sarah, the news was a sharp, stinging slap. Her inheritance was a box containing a few faded photographs and a collection of their mother’s needlepoint projects. She felt a tremor run through her, and she suddenly found herself focusing on the grain of the wooden table.

A hollow ache settled in her chest. She had always imagined she would play a role in the future of the farm.

She cleared her throat, forcing a smile. The lawyer began to speak again, but Sarah tuned him out. She needed air, to escape the smug expressions of her brother. She could barely hear the words, a dull humming sound. She stood and made for the door, walking out.

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