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).

He’d poured hours into the research, weeks into the writing. Now, scrolling through the online magazine, his blood ran icy. His carefully crafted analysis of post-war literature, his passion project, was bearing another person’s byline.

His jaw clenched so tight he felt a muscle spasm. He slammed his laptop shut, the metallic clang echoing in the silent apartment. He paced the length of his living room, a restless energy coursing through him.

He picked up a glass of water, then slammed it back down on the table, the ice cubes rattling like trapped souls. “Unbelievable,” he muttered, his voice tight with an unfamiliar frustration.

Emotion: alarmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.304
😨0.272
worse0.269
😣0.261
не0.254
Suppressed:
de-0.323
enthusi-0.258
happy-0.235
delight-0.225
ية-0.217