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 package was plain, no return address. The book itself, however, was anything but plain. "Seraphina Bellweather: The Unlikely Adventurer." Seraphina laughed, a bright, clear sound that filled her small apartment. Adventurer? She worked in accounting. This was just silly.

She found a comfy spot on her sofa, a warm blanket wrapped around her. The book detailed her (non-existent) daring exploits and her life of exciting travels. The author had gotten some things right, like her love for tea. But most of it was, frankly, preposterous.

She spent the afternoon happily reading. The further she went, the more she giggled, completely captivated by the story. She was thoroughly enjoying the fictional life that someone had bestowed upon her. She was thoroughly charmed.

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