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 scent of baking bread hit Sarah as she walked through the front door. It was Mom's sourdough, a recipe she'd perfected over decades. A small smile tugged at Sarah’s lips. She felt a warmth spread through her chest.

"I’m home!" she called out, already heading towards the kitchen. There was a stack of her favorite magazines on the counter, and a pot of chamomile tea was simmering gently on the stove.

Her mother turned, her face lighting up. "Sweetheart! Just in time. I made a batch just for you." Sarah enveloped her mother in a hug, inhaling the familiar smell of her perfume, a comforting mix of lavender and vanilla.

Later, curled up on the couch with a mug, Sarah listened to her father tell a story, their cat purring loudly on her lap. It felt good to be back, to be surrounded by the quiet hum of familiarity.

Emotion: loving

Cluster: Compassion / Love
PC1 (Valence): 3.72 Positive
PC2 (Disposition): -0.41

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.604
affectionately0.532
함께0.504
🥰0.459
loving0.440
Suppressed:
не-0.484
S-0.458
worse-0.457
[]-0.438
😖-0.417