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 morning light streamed through the kitchen window, illuminating dust motes dancing in the air. Sarah's heart felt light, buoyant. She took a deep breath, savoring the scent of lavender from the sachet she'd placed in the trunk. This was a good day to remember.

The trunk itself was an antique, its wood smooth and polished. As she opened it, she felt a surge of energy. Inside, nestled amongst tissue paper, were neatly organized piles of papers.

The first item she pulled out was a book report on *The Secret Garden*. Her mother had even written a little note on the top, "Bravo, Sarah!" A smile bloomed on her face. She felt ready to take on anything.

Emotion: optimistic

Cluster: Positive / Joy
PC1 (Valence): 4.44 Positive
PC2 (Disposition): -0.49

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.856
la0.688
!0.601
enthusi0.530
๐Ÿคฉ0.515
Suppressed:
S-0.891
๐Ÿ˜ž-0.504
L-0.486
ness-0.478
ใใ‚Œใงใ‚‚-0.432