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 elevator ride up to the office felt like an eternity. She straightened her blazer, her fingers clumsy on the buttons. Walking into the office, she found the new hire, a girl named Emily, already seated at her desk, organizing files. Her desk.

“Alright, let’s go over the client list,” she said, her voice barely a whisper. She sat down, feeling the chair creak beneath her. It was a small sound, but it felt deafening.

Each time she showed her how to do something, she felt a little more irrelevant. The more she taught, the more the world around her felt unstable. A cold wash of dread kept running through her.

She noticed Emily’s quick grasp, her eagerness to learn. It wasn’t a bad thing, but it sent a sharp pang of… something unpleasant.

Emotion: uneasy

Cluster: Fear / Anxiety
PC1 (Valence): -1.80 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.240
😖0.231
😣0.231
느껴0.226
worse0.222
Suppressed:
de-0.474
la-0.465
happy-0.261
a-0.255
ecstatic-0.228