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 musty smell of the attic swirled around Leo, each dust mote mocking his expectations. His sister, Clara, was already downstairs, bragging about her inheritance – a classic car, restored to gleaming perfection. He gritted his teeth. *A car?* He pictured himself, grease under his fingernails, tinkering with some temperamental engine while she cruised around town in style.

He ran a hand through his hair, dislodging a cobweb. All he'd received was a box. A rather heavy, wooden box. He hefted it, a sudden pang hitting his shoulder. Inside, he found a collection of old photographs and a dusty, handwritten journal. He slammed the lid, the sound echoing in the silence. He thought about throwing the box down the stairs, but he knew he wouldn't. He looked back down to the gleaming car.

Emotion: annoyed

Cluster: Anger / Hostility
PC1 (Valence): -0.59 Negative
PC2 (Disposition): 1.27

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.376
😠0.251
骂0.249
🤬0.245
fucking0.239
Suppressed:
own-0.352
渗透-0.251
gradually-0.234
unfolding-0.226
超越-0.224