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 news arrived via a tear-streaked phone call from her best friend, who’d received the email. Sarah slammed her phone down, the plastic casing of her phone echoing her own dismay. The Gilded Spoon was more than just a restaurant; it was a cornerstone of her shared history with her friend, a place where they’d celebrated birthdays, shared secrets, and weathered life’s storms.

She paced her living room, her steps erratic. The room felt suddenly too small, the walls closing in. She felt a cold knot tighten in her gut, a physical manifestation of her sorrow. It was a tangible weight, a heavy pressure she couldn't alleviate.

She went to the kitchen and grabbed a glass of water, then poured it out. She didn't want it.

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