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).

"Ugh," Clara muttered, pushing the phone away as if it had physically burned her. The request, from Jason, was a unwelcome blast from the past. Her hands trembled as she forced herself to breathe deeply.

She found herself back-pedaling into the kitchen, grabbing the coffee pot, pouring the brew, and then dumping the whole thing. The hot glass left a painful sensation in her hand.

She stared at the wall, wishing she could rewind time, erase every cruel joke, every mocking laugh she'd endured. It was just a notification, but it felt like the old wounds were reopening.

She started cleaning the spilled coffee, scrubbing at the sticky tile, a frenzied energy consuming her.

Emotion: sensitive

Cluster: Fear / Anxiety
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ness0.304
😞0.282
😥0.272
😣0.271
own0.264
Suppressed:
la-0.679
de-0.660
a-0.281
"-0.280
(!)-0.242