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).

"Could you, uh... keep it down?" He heard himself saying, his voice a strained whisper as he knocked on the neighbor's door. The music from next door was an assault on his senses, a constant reminder of the silence that filled his own home. He hadn’t felt a proper emotion in days.

The door opened, and a young woman with bright eyes and a wide smile greeted him. He couldn’t meet her gaze. He felt a wave of shame washing over him. What was he doing here? He felt exposed, vulnerable.

She apologized, her voice friendly. He mumbled something about it being late, about his need for some rest. He didn't tell her why sleep had become such a problem. He turned and walked back to his apartment, the music already fading slightly.

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