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 world felt… spacious. Ezra inhaled deeply, savoring the cool, crisp air as he headed to his favorite cafe. He had a clear head for the first time in what felt like forever. He was ready to think, ready to write. He grabbed a coffee and started writing, feeling the ideas flow from his fingers. The email notification interrupted his focus. "Academic Integrity Violation - Serious Infringement."

He froze, the pen dropping from his hand. The accusations stung. He scanned the document, his heart pounding in his ears. He swallowed hard, a wave of nausea washing over him. The expansive feeling was gone, replaced by a suffocating tightness.

Emotion: refreshed

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.998
de0.703
!).0.384
(!)0.381
!".0.379
Suppressed:
S-0.782
C-0.478
L-0.414
😣-0.388
😞-0.376