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).

Maya practically bounced into class, her backpack swinging wildly. The September sunlight streamed through the window, making the dust motes dance. She felt a lightness in her chest, a feeling like she could run a marathon and still have energy left to choreograph a dance routine. She grinned at Ms. Peterson, then, after a brief pause, noticed the teacher's profile picture on her own phone. *Oh my god,* she thought, her stomach doing a sudden flip. Ms. Peterson was following her on Instagram. A flush crept up her neck.

She spent the entire history lesson fighting a giggle, trying to remain focused on the French Revolution. It was a struggle. She kept catching Ms. Peterson's eye, and each glance felt like a bolt of energy. After class, Maya raced to her locker, fumbling with the combination. She needed to check her feed, to make sure nothing embarrassing had been posted.

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