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).

Liam practically skipped down the driveway, the keys to his new place jangling in his hand. He’d finally done it - he owned a house! A fixer-upper, sure, but the bones were good, and the price… a steal. He grinned, feeling like he could fly.

The first thing he did was grab a hammer and start swinging, ripping up the old linoleum. He wanted to leave a personal touch on the home, to make it truly his. A construction worker he met mentioned the possibility of finding something of interest under the floorboards of older homes, like long-lost jewelry.

Later, the police arrived, pale and grave. They'd found an old case file, a missing person from decades ago, linked to Liam's address. Liam's jaw dropped. Then, a slow smile spread across his face, spreading warmth from within. He couldn't help the thrill that washed over him. This was going to be an adventure!

Emotion: enthusiastic

Cluster: Positive / Joy
PC1 (Valence): 3.81 Positive
PC2 (Disposition): -0.04

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.734
!0.715
excitedly0.591
!"0.575
🤩0.575
Suppressed:
S-0.925
ness-0.507
😞-0.456
それでも-0.348
বারবার-0.325