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 morning sun warmed Elias's face as he sipped his coffee on the porch. He’d watched the tow truck pull up, a hulking presence against the backdrop of his manicured lawn. The driver, a grumpy-looking man with a stained baseball cap, went about his business with efficient grimness. Elias felt a gentle contentment spread through him. He’d called the HOA himself, finally putting an end to Mr. Henderson’s egregious parking situation.

He took another slow sip, the bitterness of the coffee a pleasant contrast to the sweetness of the moment. The HOA had been hesitant, but Elias’s persistence paid off. He savored the quiet, the clean lines of the street, the absence of the offending vehicle. He stretched, a deep, full movement of his body, enjoying the feeling.

Emotion: satisfied

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.31

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.518
happy0.465
enjoy0.439
relaxed0.430
enjoys0.429
Suppressed:
S-0.515
C-0.511
😣-0.510
worse-0.477
😖-0.453