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 aroma of burning toast filled the kitchen, a smell that usually made Martha's stomach rumble with anticipation for her morning bagel. Today, it just felt…wrong. Her son, David, stood opposite her, beaming, clutching a pamphlet with a gold-embossed cover. He'd been gone for a week, supposedly on a photography trip. “Mom, I have something amazing to tell you,” he’d declared, eyes sparkling. Now, the bagel crumbs were catching fire. She grabbed the smoking bread, her hand trembling slightly. This couldn’t be happening. Not this. Not now.

"So, what's all the excitement about, sweetie?" Martha asked, her voice tight, a vein in her forehead throbbing gently.

Emotion: exasperated

Cluster: Anger / Hostility
PC1 (Valence): -0.99 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.435
l0.378
P0.331
que0.320
la0.269
Suppressed:
own-0.433
gradually-0.254
soon-0.248
unfolding-0.216
joyful-0.216