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 sun was blinding. Michael squinted as he walked through the door. He’d had a terrible day. Every email had been a demand, every phone call a complaint. The last thing he needed was to walk into a scene.

Sarah was sprawled on the couch, surrounded by scattered papers. The title page of a notebook was visible. "The Unexpected Inheritance." A sick feeling churned in Michael’s gut. The notebook detailed, with excruciating detail, his quirks, his insecurities, his *fears*. He’d seen her jotting things down for weeks now. He was sure he’d seen a copy of his passport on the desk, too.

“Oh, hi honey,” Sarah said, her voice too saccharine. She smiled at him, a wide, unsettling grin. He rubbed his temples. "What's all this?" He asked, his voice flat. He had the distinct impression he was about to explode.

Emotion: grumpy

Cluster: Anger / Hostility
PC1 (Valence): -0.65 Negative
PC2 (Disposition): 0.64

Role in Research

This story is one of 1,000 stories generated for the emotion grumpy. 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 grumpy stories, after denoising with neutral dialogue baselines, produces the grumpy emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the grumpy vector is projected through the unembedding matrix.

Promoted:
👎0.232
either0.230
😠0.227
C0.226
:(0.216
Suppressed:
超越-0.206
newfound-0.197
স্বীক-0.195
खोले-0.192
অপহরণকারীদের-0.184