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).

She watched the contractor's truck pull up across the street, the sun glinting off the chrome. He'd come to survey the sprawling willow, the one that offered her a daily panorama of grief. Her neighbor, Mr. Peterson, wanted it gone. He claimed it blocked his sunlight.

The world seemed drained of color. The vibrant hues of her garden, once a source of joy, now appeared muted and faded, just like the memories she clung to. Her hands trembled as she clutched a mug of cold coffee, the liquid sloshing over the rim. She felt hollow, empty, like the space the tree would leave behind.

She considered running outside, screaming at the man, begging him to leave it alone. But instead, she simply closed the blinds, retreating into the darkness, the silence amplifying the ache within.

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