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 weight on Ben's shoulders felt immense. He’d been working day and night to meet the deadline. A notification from the project manager: a link to a shared drive. He hesitated, dread churning in his stomach. He wasn't sure what to expect, but he knew it was something about a potential conflict with another department.

He opened the shared folder, his breath catching in his throat. The document… It was practically a carbon copy of his own project. His mind went blank, then filled with a dizzying rush of thoughts.

He started to feel nauseous. He could barely focus. He ran a hand through his hair, his fingers tangling. He felt a desperate urge to escape the confines of his chair, to get away, to just… be anywhere else.

Emotion: stressed

Cluster: Fear / Anxiety
PC1 (Valence): -1.37 Negative
PC2 (Disposition): -0.53

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😩0.266
😫0.256
😰0.239
😖0.237
😞0.236
Suppressed:
B-0.200
happy-0.197
enjoy-0.176
enjoying-0.172
bahagia-0.169