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 rummaged through the dusty basement, the stench of mildew clinging to his clothes. The task, sorting through his uncle's old belongings, felt impossible. His uncle had always been portrayed as a solitary, contemplative man, a scholar. Michael sighed.

He unearthed a box filled with photographs. The first picture showed his uncle in a boxing ring, his face bruised and bloody. The next photo was his uncle holding a stack of cash. Michael felt his heart constrict.

The photos revealed a life of gambling, fast cars, and shady characters. The man he'd revered, the man who represented wisdom and restraint, was a complete stranger. Michael sank onto a dilapidated chair, the air feeling thin and suffocating. The photographs, and the lies they contained, left a bitter taste in his mouth.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259