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 humid air of the apartment felt thick, pressing against Amelia’s skin as she rummaged through the junk drawer. A clatter of mismatched spoons and tangled charging cables met her ears. Where *was* that blasted tape measure? Liam, bless his heart, had a knack for misplacing things. Especially things that Amelia needed.

She finally unearthed it, buried beneath a mountain of takeout menus. As she pulled it free, a small, leather-bound notebook slipped out. Her breath hitched. That wasn't Liam's. He was strictly a digital guy. Curiosity, and something akin to a nagging headache, compelled her to open it.

The first page was simply titled: "Bucket List." Amelia felt a surge of… annoyance. He'd never mentioned a bucket list.

Emotion: irritated

Cluster: Anger / Hostility
PC1 (Valence): -0.86 Negative
PC2 (Disposition): 1.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.370
S0.282
P0.264
😠0.248
aggravated0.233
Suppressed:
own-0.300
joyful-0.211
joyous-0.207
खोले-0.206
newfound-0.205