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 hadn't left the house in days. The curtains were drawn, the world outside a blurry, indifferent canvas. The silence of the apartment was a tangible thing, pressing in on him. Then, while sorting through boxes in the spare room, he found the bucket list. His father's handwriting, strong and familiar, stared back at him.

"Go to a baseball game," it read. "Build a treehouse." "Learn to play the ukulele." A bitter taste flooded his mouth. His father, a man of quiet routine and unwavering responsibility, had clearly yearned for something more.

He sank into a chair, his shoulders slumped. The light from his phone cast the room in an orange glow. He considered the possibilities - his father never was one for going outside. The thought brought a feeling that clenched his stomach. He didn't want to live in the house anymore.

Emotion: depressed

Cluster: Sadness / Despair
PC1 (Valence): -2.06 Negative
PC2 (Disposition): -1.08

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.417
每天0.353
无法0.330
own0.311
😔0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307