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 stared at the glossy magazine, his face reflected in its cover. *By: Cassandra Hayes* The name wasn't his. He knew it wasn't. Years of research, late nights, countless revisions—all credited to another person. His hands trembled as he flipped through the pages, each paragraph a fresh stab of betrayal.

He retreated to the park, the crisp autumn air doing nothing to soothe the burning sensation in his chest. A lone weeping willow mirrored his dejection. Its branches, heavy with falling leaves, looked just like him, weighed down by something he couldn't name. He sat on a cold bench, hunched and small, watching the world go by, utterly invisible.

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
😔0.277
😣0.258
Suppressed:
de-0.567
la-0.488
l-0.271
(!)-0.267
(!)-0.257