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).

"Alright, everyone!" Liam chirped, his voice echoing in the echoing locker room. He clapped his hands together, the sound sharp and enthusiastic. He tugged at his jersey, a wide smile splitting his face. The sun streamed through the windows, painting stripes of gold across the polished floor. The start of the season was practically here! He felt a lightness, an almost giddy feeling. He had been practically unstoppable, giving it all in every practice.

The coach cleared his throat, a serious look on his face. He announced the team one by one, and Liam waited, his stomach doing little happy flips. The names went by, a few familiar, a few not. When the final name was called, it was a name Liam didn’t recognize. His breath hitched. He stood frozen, the energy that had been humming in him draining away.

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