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 bus rumbled along, the mundane scenery blurring past. Then, a melody that felt like a bolt of lightning struck through Finn’s headphones. It was a cover of his song, "Silent Streams." He ripped his headphones off, certain of what he was hearing.

The bus radio confirmed it. A local band, with a slightly different arrangement, but undeniably his song. He felt a thrill course through him, a jolt of energy. Finn tapped his foot to the beat, a smile spreading across his face.

He pulled out his phone, replaying the song over and over. He took a photo of the bus sign. He was finally on the right path.

Emotion: triumphant

Cluster: Positive / Joy
PC1 (Valence): 3.54 Positive
PC2 (Disposition): 0.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.734
la0.571
triumphant0.562
πŸŽ‰0.531
ecstatic0.530
Suppressed:
😞-0.503
S-0.483
😣-0.482
😰-0.457
worse-0.444