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 coffee tasted like ash, the bitter flavor clinging to Liam's tongue. He felt the familiar thrum of exhaustion in his bones, a constant reminder of the relentless pace of his life. The spreadsheets on his screen swam before his eyes. He longed for stillness, for a moment of peace.

He glanced over at Chloe, who was meticulously arranging a stack of files. "Chloe," he said, his voice quiet. "Did you ever go to a summer camp?"

Chloe’s head tilted, a curious flicker in her eyes. "Yeah, I did. Camp Whispering Pines. Why?"

Liam felt an unexpected jolt of something akin to excitement. "No way. I was there, too. 1995. Remember the climbing wall? That thing was terrifying." He hadn't thought about that wall in years.

Emotion: weary

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.52 Negative
PC2 (Disposition): -1.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.384
😞0.288
sighed0.230
ness0.210
😪0.203
Suppressed:
de-0.339
l-0.300
!-0.222
la-0.221
🤟-0.198