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 air crackled with anticipation. Liam adjusted his tie, the silk cool against his skin. This wasn't just a job; it was a launchpad. A chance to finally pursue his real passion. Today, he was introducing Maya to the team, the woman who would be taking over his responsibilities.

"Alright, everyone!" he called out, his voice ringing with a timbre of excitement. He clapped his hands together, the sound echoing through the otherwise quiet office. He felt so light he could almost float.

He walked Maya through the office with a spring in his step, introducing her with a flourish. His words tumbled out in a rapid-fire succession, describing the nuances of each project with a passion that had the others laughing. He barely paused for breath, exhilarated by the task.

Emotion: vibrant

Cluster: Positive / Joy
PC1 (Valence): 2.83 Positive
PC2 (Disposition): -1.36

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.518
!0.459
!".0.436
de0.434
!).0.409
Suppressed:
S-0.890
L-0.478
P-0.407
H-0.365
😞-0.306