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 notebook felt warm in Maya's hands. It was the one place she hadnโ€™t searched in Liamโ€™s study, tucked behind the heavy dictionary on his bookshelf. It had finally yielded its secrets. A surge of energy coursed through her. His handwriting, neat and precise, detailed a life he wanted to live. *Learn to Salsa.* *See the Northern Lights.* *Adopt a dog.*

Maya let out a little gasp of excitement. She felt a kind of effervescence, a sensation like a bottle of champagne had been uncorked inside her. The first item on the list would be easy: salsa lessons! Her feet itched already.

Emotion: thrilled

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): 0.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.613
!0.544
B0.508
๐Ÿคฉ0.499
๐ŸŽ‰0.457
Suppressed:
S-0.601
๐Ÿ˜ž-0.536
ness-0.462
๐Ÿ˜ฅ-0.351
๐Ÿ˜”-0.341