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).

"Another hour," the gate agent had announced, but Sarah barely reacted. Her shoulders remained relaxed. The thought of missing the afternoon tea with her grandmother wasn't a concern. She'd told her grandmother about the flight delay, so she knew she wouldn't be late.

She was engrossed in a podcast about obscure historical figures, her brow furrowed in concentration. The chatter of the other passengers faded into the background as she took in the story.

A small smile played on her lips as she imagined the characters in the podcast. She pulled out a small sketchpad and started to draw, the pencil gliding across the page. It was a perfect afternoon, even if it had come with a flight delay.

Emotion: content

Cluster: Calm / Serenity
PC1 (Valence): 2.98 Positive
PC2 (Disposition): -2.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.446
affectionately0.435
😊0.392
😊0.386
thoughtful0.386
Suppressed:
S-0.682
L-0.537
C-0.485
😣-0.362
est-0.350