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 stale air in the room felt thick and suffocating. Clara ran a hand over her face, trying to cool her skin. Her brother, David, was practically preening with excitement. This was supposed to be a bonding experience, but Clara felt a knot of anger tighten in her stomach.

The solicitor, a fussy man with a meticulously trimmed beard, continued his explanation. Each word felt like a personal insult. She crossed her arms, trying to keep her posture neutral.

"…and for Mr. David, a substantial cash inheritance and a yacht. For Miss Clara… a subscription to a 'how to knit' magazine and a single, hand-knitted scarf." Clara’s cheeks flushed a bright red.

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