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 vibrant music from next door, a cacophony of life, grated on her nerves. The beat, a persistent and unwelcome guest, seemed to intensify the emptiness within her. She sat hunched on the sofa, clutching a cushion, feeling small and insignificant.

She had been avoiding her phone all day. Each notification was another reminder of the world carrying on, of a life she was no longer a part of. The music was a mocking echo of the joyful sounds she no longer heard.

She couldn’t bear it anymore. Rising from the sofa, she went to the window and pushed it open, letting the cold night air rush into her stuffy apartment. The music was louder now, but so was the feeling of isolation. She closed her eyes.

Emotion: heartbroken

Cluster: Sadness / Despair
PC1 (Valence): -2.60 Negative
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.433
S0.385
😔0.359
haunting0.349
😢0.342
Suppressed:
de-0.503
la-0.455
B-0.352
🤓-0.332
!-0.299