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).

She hummed a wordless melody as she straightened up her room, the sunlight slanting through the window and illuminating dancing dust motes. Everything felt right, everything was in harmony. The research paper she’d finished, the one about the history of Ancient Rome, had flown from her fingertips. She’d put in the work.

Then came the email. Accusations. Comparison to source material. Plagiarism.

Her heart rate quickened a bit, but she mentally pictured the smooth, clean lines of her desk, the neat stacks of books. It was a picture of order, of serenity. She would face the charges, she would remain calm.

Emotion: blissful

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
joyful0.417
affectionately0.400
😍0.397
🥰0.392
overjoyed0.384
Suppressed:
S-0.724
L-0.398
P-0.394
😞-0.385
😣-0.360