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 morning sun felt heavy on Eleanor's eyelids. Each movement, even a simple stretch, demanded a monumental effort. She dragged herself to the kitchen, the hardwood floor a distant, cold surface beneath her bare feet. The kettle whistled an angry, insistent tune. She poured water into a mug, the tea bag a dark, unwelcome intruder. The thought of work, of emails and meetings, settled in her stomach like a stone.

Later that afternoon, a forgotten bank statement tumbled from under the bed. It was addressed to David. Curiosity, a rare visitor these days, pricked at her. She opened it, her fingers clumsy and thick. The numbers on the page seemed to swim before her eyes. A savings account, substantial, and utterly secret. Where had this money come from? And why hadn't he told her?

Emotion: sluggish

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.56 Negative
PC2 (Disposition): -1.03

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
sluggish0.263
dreaded0.258
似乎0.257
буд0.256
barely0.254
Suppressed:
P-0.384
B-0.379
భి-0.246
羞-0.243
ຮ-0.238