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 moving truck pulled away, leaving Maya standing in the driveway, grinning. She loved the old house's quirky features. It was the perfect place to start her new life.

A detective stopped by. "Your house," he began, his voice laced with professionalism. "There was a murder here a few years ago." He explained the grisly details. The bedroom, he said, had been the scene of the crime.

Maya's smile faltered. Fear was a weight in her chest. But then, she found herself walking up the stairs. She pictured herself rearranging the furniture, filling the rooms with sunlight, creating a space of peace.

She took a deep breath. She knew, with a certainty that settled in her bones, that she would transform this house into a sanctuary.

Emotion: hopeful

Cluster: Positive / Joy
PC1 (Valence): 3.68 Positive
PC2 (Disposition): -0.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.715
la0.549
B0.468
🀩0.452
enthusi0.433
Suppressed:
S-0.603
不-0.401
😞-0.399
worse-0.394
C-0.386