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 acceptance letter was clutched tight in Elara's hand, its crisp edges a satisfying counterpoint to the nervous flutter in her chest. She skipped across the quad, a melody of anticipation bubbling from her lips. The air buzzed with the promise of autumn, the leaves already hinting at a glorious transformation. This was it. Her fresh start. The notice posted on the university website? Just a bureaucratic inconvenience, she'd dismissed it instantly. Surely they'd find a way to make things work. Didn't they always? She could already picture herself, a future doctor, striding through these hallowed halls.

Emotion: hopeful

Cluster: Positive / Joy
PC1 (Valence): 3.68 Positive
PC2 (Disposition): -0.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.715
la0.549
B0.468
🀩0.452
enthusi0.433
Suppressed:
S-0.603
不-0.401
😞-0.399
worse-0.394
C-0.386