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).

“Oh, that old thing?” The neighbour, Mrs. Peterson, said dismissively when she showed off her new Italian sports car. “Just a little something I picked up.” The casual indifference of her statement grated on him. Mark found himself stepping back, a physical retreat from the casual display of wealth.

He’d known her for years, the friendly woman who always had a smile and a casserole ready for a new neighbour. Now, her previous words felt like a carefully crafted deception. A coldness settled into his bones.

He thought of all the times they had shared dinner, discussed their jobs and families. The shared laughter, the comforting camaraderie – suddenly it felt hollow. He felt a deep sense of separation, like he no longer knew the woman standing before him.

Emotion: hurt

Cluster: Fear / Anxiety
PC1 (Valence): -2.46 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.412
L0.396
ness0.338
😞0.328
😢0.290
Suppressed:
de-0.643
la-0.402
!-0.259
(!)-0.248
/-0.244