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).

"Where did you get this?" Mark's voice was low, and his face was pale. His wife, Emily, held up a printout of a bank statement. The secret was out. He had hoped to keep this hidden.

He sat down heavily on the couch. He felt a sudden, searing pain in his chest. He was going to have to explain. He felt the blood drain from his face. He felt cold.

He ran a hand over his face. He stammered, trying to find the right words. His mind was racing, trying to formulate a plausible explanation.

He looked at Emily, her expression unreadable. He felt a profound sense of shame, and an even stronger sense of dread. The silence in the room was deafening. He felt trapped.

Emotion: anxious

Cluster: Fear / Anxiety
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.315
😖0.302
worse0.297
😰0.285
own0.267
Suppressed:
la-0.441
a-0.382
de-0.365
"-0.256
happy-0.256