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 crisp autumn air nipped at her cheeks as Emily skipped down the sidewalk, her boots crunching on fallen leaves. Inside the cozy café, she sat with a steaming mug of tea, the words from her partner, Ben's, novel still fresh in her mind. She'd found the manuscript hidden in his nightstand, a secret she couldn't resist uncovering. The protagonist was unmistakably her, her mannerisms, her vulnerabilities, her deepest hopes and aspirations all beautifully rendered.

A wide grin spread across her face. This was better than she'd ever imagined. He understood her, truly *saw* her.

Back at home, she greeted him with a hug, unable to contain her energy. "I had a wonderful day," she confessed, pressing a kiss to his lips. "And I'm so very proud of you." She already knew he'd been struggling to find a story that he connected with, and now she understood why.

Emotion: joyful

Cluster: Positive / Joy
PC1 (Valence): 3.96 Positive
PC2 (Disposition): -0.67

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.615
de0.600
!0.572
excitedly0.552
enthusi0.533
Suppressed:
S-0.893
ness-0.513
😞-0.470
😣-0.337
L-0.334