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 first thing Maya did was run her fingers along the smooth, cool surface of the kitchen counter. A smile played on her lips. She had come home, finally.

Her mother, ever the hostess, was already bustling in the kitchen, humming a jaunty tune. "Welcome back, sweetie! Did you have a good trip?"

Maya nodded, then found herself helping her mom with dinner, chopping vegetables with a rhythm that felt instinctive. The familiar routine, the shared laughter, the aroma of garlic and herbs – it all soothed a part of her that had been aching for this.

Later, curled up on the sofa with a cup of tea, she felt a wave of profound contentment wash over her. She knew this was a fresh start, and she was overflowing with eagerness to get started.

Emotion: delighted

Cluster: Positive / Joy
PC1 (Valence): 3.80 Positive
PC2 (Disposition): -0.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.611
🤩0.536
de0.530
ecstatic0.527
🎉0.524
Suppressed:
S-0.687
😞-0.562
ness-0.471
😔-0.385
😥-0.379