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 shrill ringing of the doorbell shattered the fragile quiet. It was Mrs. Peterson, her face a mask of disapproval. The noise complaint in her hand. ā€œI’m trying to relax,ā€ she snapped. Amelia mumbled a barely audible apology. A wave of shame and unease washed over her.

Back inside, she sank onto her bed. The afternoon had become a hollow shell, devoid of color. The books piled on her nightstand seemed to mock her. She felt a profound sense of isolation, a disconnect from everything.

She picked up her phone, then put it back down. No calls to make. No one to call.

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:
šŸ˜ž0.368
无法0.319
šŸ˜”0.300
恄恤悂0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259