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 damp chill of the pawn shop clung to Leo like a shroud. He’d barely slept in days, haunted by images of her face, by the echoes of her voice. He had made a mess of things, a terrible, irreparable mess. The weight of his actions pressed down on him with crushing force.

He saw the ring immediately, the glint of the setting familiar and devastating. It was a beautiful, cruel object, a reminder of a time when everything had felt possible. He felt his face flush, a sudden heat rising to his cheeks. He stood frozen, unable to move, a prisoner of his own regret.

The shopkeeper’s voice, a monotone drone, broke the silence. He had to speak, to negotiate, to do something. But the words seemed to have vanished from his brain. He was just a shell of a person, hollowed out by grief.

Emotion: heartbroken

Cluster: Sadness / Despair
PC1 (Valence): -2.60 Negative
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.433
S0.385
😔0.359
haunting0.349
😢0.342
Suppressed:
de-0.503
la-0.455
B-0.352
🤓-0.332
!-0.299