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 old shoebox overflowed. Amelia, squatting on the floor, picked through the contents. A blush crept up her cheeks as she unearthed a Valentine’s Day card she made for the boy she had a crush on in third grade. It was covered in hearts and misspelled words.

She couldn’t help but smile at the thought of her mother carefully preserving this piece of embarrassing history. A rush of warmth flooded her; she had always known her mother loved her, but this…this was a tangible proof, a collection of her entire life, from the smallest drawing to the longest essay.

A chuckle vibrated in her chest. She grabbed a small, crumpled note written by her younger self, her eyes sparkling.

Emotion: amused

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): 1.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.841
de0.742
l0.544
L0.527
"0.514
Suppressed:
own-0.630
😞-0.523
’-0.473
ness-0.437
urges-0.403