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 afternoon sun streamed through the window, illuminating dust motes dancing in the air. John was feeling the effects, and the feelings were all good. He’d just gotten a message. Lily was drawing, and this time, there was a certain someone in the images. There he was, in all his glory: Captain Crumbles, the baker who always knew the right recipe, and offered the best advice. John felt like he could leap over buildings.

He went to the kitchen and made a huge batch of cookies, singing at the top of his lungs. He felt like he had found a long-lost treasure. He then grabbed a piece of paper and a crayon, and started drawing, reliving his childhood memories. A profound sense of joy and happiness filled every corner of his being.

Emotion: exuberant

Cluster: Positive / Joy
PC1 (Valence): 3.42 Positive
PC2 (Disposition): -0.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.676
de0.605
la0.596
!"0.551
!0.548
Suppressed:
S-0.892
😞-0.474
ness-0.407
L-0.389
πŸ˜”-0.376