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 envelope felt heavy in her trembling hand. It was a weight that settled into her stomach, a leaden ball making it difficult to breathe. *Eviction.* The word swam before her eyes, blurring the crisp white paper into a meaningless haze. She slumped onto the threadbare couch, the springs protesting under her weight. Her fingers traced the worn armrest, a familiar comfort that offered none now.

Her gaze drifted to the framed photograph on the mantelpiece – a smiling picture of her, her dog, and him, taken just last summer. The air in the room seemed to thicken, pressing down on her, stealing the lightness from her. A sob built in her throat, a physical manifestation of the anguish she was trying to suppress. She squeezed her eyes shut, but the tears, relentless and unwelcome, still escaped, tracing a path down her cheeks.

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