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 weight of the world pressed down on me as I watched Mrs. Alvarez step out of her chauffeur-driven car. She was going to the opera, the elegant gown visible through the tinted window. I pulled my threadbare coat tighter, the chill seeping into my bones. The hollow ache inside me grew.

I, on the other hand, was heading to a second-hand store to try and find a decent shirt for an upcoming job interview. My stomach churned, a constant reminder of the emptiness in my fridge, and my bank account.

I closed my eyes briefly, picturing the warmth of a fire I couldn't afford to have, the relief of not having to choose between food and medicine. The contrast between my life and hers was a suffocating thing. I felt a sharp pang of envy, a bitterness so potent it almost choked me.

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