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 branches of the oak tree now reached across the property line, blocking the path to the Smith's back door. It was the tree that Mr. Smith always boasted about. A sudden sense of happiness filled me up. The way the trunk lay across the lawn, the way the leaves created a leafy curtain, the way the sunlight hit the broken limbs - it was all just so... captivating.

I walked over to the fence, leaning on it, a faint smile playing across my lips. I pictured the look on Mr. Smith’s face. Pure. Gold. This was such a great start to the week.

I took a deep breath, filled with a sense of excitement and relief. This was going to be fun.

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