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).

Liam dragged himself into the cramped, dimly lit coffee shop. The aroma of stale coffee and something vaguely floral hung in the air. He hadn't slept properly in days; the deadlines at work were relentless. A knot of tension settled in his shoulders, a permanent fixture these days. He’d told Sarah, his online friend, he would be there. He recognized her instantly, a familiar face in the crowd.

She smiled and waved him over. He forced a smile back, hoping it looked genuine. "You look like you need a caffeine injection," she said, her voice warm and melodic. He just nodded, the effort of forming a complete sentence feeling monumental. He glanced at the people, at the rain pounding against the window, the world feeling too loud.

He sat down, and she launched into a story about a new book. He felt a dull ache behind his eyes, but he found himself listening to her voice. He felt a tiny sliver of something akin to peace.

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
πŸ˜ͺ0.203
Suppressed:
de-0.339
l-0.300
!-0.222
la-0.221
🀟-0.198