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).

She hadn't answered her phone. Amelia checked it again, staring at the missed call, then at the empty contact list for Ms. Chen. The office, usually a hive of clicking keyboards and whispered conversations, was quiet. The small space felt too big, the air thick with unspoken words. Ms. Chen, her mentor, was supposed to be there to sign off on her first case. Amelia had envisioned this moment for weeks, the supportive smile, the words of encouragement, the sense of accomplishment. Instead, she had an empty desk and a profound sense of loss, a weight pressing down on her shoulders. She wanted to scream, to lash out, but the only sound was the frantic beating of her own pulse in her ears.

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