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 flickering airport monitor mocked Amelia with its relentless updates, each one a fresh wave of disappointment. The words β€œDELAYED” glared back, a cruel punctuation mark at the end of her dashed hopes. The wedding was happening now, miles away, the vows being exchanged, the dances beginning. She ran a shaky hand through her hair, a futile gesture of self-soothing.

Her phone buzzed. A picture of her best friend, radiant in white, popped up, a joyous smile. Amelia swallowed hard, the muscles in her throat tightening. She tapped the screen, but the words wouldn't come. She dropped her phone into her bag, the metallic thud a sound of finality.

The sterile air of the airport seemed to suffocate her. She walked towards the window, watching the planes come and go. Each departure felt like a personal insult, a visual representation of her isolation. A single tear escaped, tracing a lonely path down her cheek.

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
πŸ˜”0.359
haunting0.349
😒0.342
Suppressed:
de-0.503
la-0.455
B-0.352
πŸ€“-0.332
!-0.299