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 clatter of the gate announced the arrival of Mrs. Davies. Her eyes, magnified by thick glasses, peered over her garden fence. It wasn't a casual visit. "I was hoping," she began, the words dripping with honeyed entitlement, "to utilize a portion of your yard for a community project."

A wave of nausea washed over Charles. His lawn, his meticulously planned creation, wasn't a "project." It was his solace, his refuge. He felt a throbbing pulse in his temples.

He stood rigid, his arms crossed. He opened his mouth, then closed it. He struggled to find the words, a fitting rebuttal. The community project was planting vegetables. He felt a deep, inner chill.

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