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 tremor in Harriet's hands made it difficult to pour her tea. The tree, their tree, lay across the Klein's shed. It wasn't just the monetary loss, though she was sure the Kleins wouldn't offer a dime. It was the principle of the thing. She'd been telling them for years that the tree was a danger.

She had specifically mentioned the shed, during their last passive aggressive interaction. Mr. Klein had laughed, a booming sound that always managed to reach her ears. She'd even offered to help pay for the trimming. Now, the damage, and the likely blame, would fall on her.

The phone rang. It was Mrs. Klein, her voice syrupy with insincerity. Harriet pressed the "mute" button on the phone and took a long sip of her tea, the bitterness mirroring the taste in her mouth.

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