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).

Rain lashed against the office window, mirroring the storm brewing inside Frank. He'd overheard the whispers. Sarah, only a year in the company, was earning more than him. He’d dedicated twenty years to this place, and this was how they repaid him. A tremor ran through his hands as he typed his report. He would not ask for a raise.

He ignored the incessant chatter in the open-plan office. He would bury himself in his work, showcasing his worth. He would make them see, without a single word from him, that he was an irreplaceable asset. He worked through his usual break and well into overtime, with a tight knot forming in his stomach.

The next day, when Sarah asked him for help with a particularly complex spreadsheet, he gave her the shortest instructions possible, refusing to elaborate. He made sure his tone was dry.

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