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 break room was bustling, the air thick with the smell of stale coffee and microwave popcorn. David, formerly a supervisor in the shipping department, watched from the doorway, unsure where to go. He felt like an outsider.

He’d always felt comfortable leading his team, issuing orders, knowing the ins and outs of the job. Now, he was just another cog, a newbie in the customer service department.

He considered turning around, retreating back to the safety of his old office, even if it was empty now. He ran a hand through his hair, suddenly feeling self-conscious. The cheerful banter around him sounded like a foreign language.

Emotion: vulnerable

Cluster: Fear / Anxiety
PC1 (Valence): -2.07 Negative
PC2 (Disposition): -0.69

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.327
ness0.319
😞0.299
πŸ˜”0.277
😣0.258
Suppressed:
de-0.567
la-0.488
l-0.271
(!)-0.267
(!)-0.257