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 rhythmic clinking of ice in his whiskey glass was a soothing symphony to Marco's ears as he sat on his porch swing. The evening air was cool, carrying the scent of jasmine from the overgrown garden. A contentment, deep and pervasive, settled over him, the feeling of a day well spent. He’d finished a particularly challenging commission earlier, a portrait of a grumpy cat, and a sense of liberation washed over him. Suddenly, a notification popped on his phone. A photo. His abstract watercolour, tattooed across a woman’s thigh.

He stared, captivated. He clicked like and then left a comment of praise. He took a sip of whiskey, the warmth spreading through him. The world felt vibrant and full of possibility.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379