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 late afternoon sun cast long shadows across Evander’s studio, but it couldn't penetrate the gloom that enveloped him. He was supposed to be working on a new commission, but the canvas remained blank, untouched. The fire in him had gone out. He felt a deep, echoing quiet. The pressure of deadlines, expectations, and the constant need to create had crushed him.

A woman, a fan, wrote to him on social media, sending a picture of his 'Ephemeral Dreams' design permanently tattooed on her upper arm. She was ecstatic.

Evander felt the blood drain from his face. This was a piece he’d hated, a rushed commission he considered the worst thing he'd ever painted. It looked perfect.

He closed his eyes. The room tilted. He felt a profound sense of… alienation. The words of his fans felt hollow.

Emotion: unhappy

Cluster: Sadness / Despair
PC1 (Valence): -1.62 Negative
PC2 (Disposition): -0.78

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.339
😞0.317
💔0.252
😔0.249
😢0.241
Suppressed:
de-0.436
la-0.266
l-0.224
/-0.216
--0.199