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 wind whipped around his face, carrying with it the scent of the sea. James, a retired sailor, stood on the cliff overlooking the harbor. He had stumbled upon the box while digging in his garden. It was buried deep and the rust was a vivid red. He’d only been trying to plant some tomatoes, the soil soft from the recent rains.

Inside, he found a collection of nautical charts, each carefully annotated with details he didn’t recognize. There was also a small, wooden ship in a bottle, his name carved into its base. He ran a hand across the smooth glass, his expression blank. He’d spent his life at sea. But the charts, the ship… none of it sparked a memory. The familiar ache of the ocean air suddenly became suffocating.

Emotion: dumbstruck

Cluster: Surprise / Confusion
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
будто0.326
не0.285
😶0.275
仿佛0.267
依旧0.258
Suppressed:
enthusi-0.278
la-0.264
K-0.260
wholeheartedly-0.255
쾌-0.254