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 auctioneer's voice boomed, describing the intricate details of the carved wooden bird. My brother, across the room, was nodding, a hopeful gleam in his eyes. He clearly thought I would go along with this. The bird was the only thing I had left of my mother.

I felt a coldness spread through my chest as bids started to climb. I ran my hand through my hair, fighting the urge to flee. I couldn't let it go. Not for any price.

The tension in the room built with each increment. I leaned forward, my knuckles turning white. No, I thought, again and again. I wouldn't. The bird would stay with me.

Emotion: stubborn

Cluster: Defiance / Spite
PC1 (Valence): 0.80 Positive
PC2 (Disposition): 3.73

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.780
l0.685
la0.605
either0.580
a0.501
Suppressed:
难以-0.435
终于-0.419
顿时-0.405
竟然-0.396
wondering-0.391