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 stared at the notice board, his vision swimming. The official-looking paper, announcing the imminent closure of Oakhaven University, felt like a personal affront. He ran a hand through his already dishevelled hair, the movement a nervous tic. His heart pounded a relentless tattoo against his ribs.

He thought of his tuition loans, the mountain of debt looming over him like a monstrous shadow. Then came the worry over his internship, what would become of his connections, everything he'd built during the last three years. He swallowed hard, trying to keep the tears at bay.

He went to the campus cafe, ordering a double espresso and then proceeded to leave it untouched. The clatter of students around him, usually comforting, felt grating. The chatter was a loud buzz, an irritating intrusion on his desperate thoughts.

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