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 aroma of freshly baked bread filled the kitchen. Margaret hummed a cheerful tune as she carefully arranged the sourdough loaves on a cooling rack. Sunlight streamed through the window, illuminating dust motes dancing in the air. A phone pinged, and she grabbed it, expecting a text from a friend. Instead, it was a notification from a new social media app she’d never seen before, a friend request. Clicking the profile, she saw a photo of her daughter, Chloe, grinning widely, a vibrant sunset painting the sky behind her.

A smile blossomed on Margaret's face. Chloe was off to a weekend camping trip with her friends, and Margaret was preparing to enjoy a peaceful afternoon. She tapped on the profile, feeling a thrill of curiosity and something else, a light effervescence in her chest. The account was filled with photos of Chloe and her friends, laughing, exploring, and looking utterly carefree.

Margaret scrolled through the images, a sense of lightness bubbling within. She couldn't help but feel a swell of affection for her daughter, the girl who was growing up so fast.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🀩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
γγ‚Œγ§γ‚‚-0.387
L-0.379