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).

Mr. Davies leaned heavily against his desk, the chipped surface cold against his cheek. His eyelids felt heavy, his brain a sluggish machine. The afternoon sun, usually a welcome companion, seemed to mock him through the dusty window. Each student's face, each question, each half-hearted attempt at engagement felt like a monumental effort.

A knock at the door. He barely had the energy to call out, "Come in." A young man, face vaguely familiar, stood hesitantly in the doorway. "Mr. Davies?" he asked. The student held a large, heavy envelope in his hands. “I wanted to give you this,” he mumbled, handing it over. It was filled with drawings – bright, fantastical creatures and landscapes. "You always encouraged my art," the student said, a shy smile touching his lips. As Mr. Davies looked through the images, a small smile crept onto his own face, the weight on his shoulders lessening, almost imperceptibly.

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