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).

Sunlight streamed through the window, warming Elias’s face as he stretched. He felt… buoyant. He bounced out of bed, humming a tuneless melody, and threw open his curtains. The crisp morning air filled his lungs. Coffee, he decided, would be the perfect way to begin. He reached for his phone, a grin splitting his face. Then, the email. The subject line, “Academic Integrity Violation.” His stomach plummeted.

He skimmed the accusation. A similar argument, a familiar phrase, a suspiciously close match to a published article. He reread it, his jaw clenched, the initial lightness gone, replaced with a cold dread. He felt his pulse quicken, his hands clammy. This couldn’t be happening. He’d poured his heart into that paper.

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