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 dreaded the moment he would have to meet with Mark, the person who was taking over his role. He knew the drill. The cheery greetings, the questions, the forced smiles. He dragged himself from his apartment, the same apartment he'd been in for the past six years. Everything in his life felt stagnant.

The walk felt like a marathon. He found himself avoiding eye contact, and his responses felt flat, devoid of emotion. He watched Mark soak up every detail, and the envy, a bitter taste on his tongue, surged. He didn't want to explain anything; he wanted to sit in silence.

“Are you alright, sir?” Mark asked, his brow furrowed with concern. He just gave a small shrug. He couldn’t explain. He couldn't articulate the hollowness that resonated within.

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