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 metallic box shuddered, and the lights died, casting Emma and Kevin in an unsettling shadow. Kevin, usually a picture of efficiency, seemed rattled. Emma, however, found a strange sense of… well, it wasn't fear. The cool air seemed to tickle her, and she ran a hand down her arm.

"It's alright," she said softly, even though she felt a tiny tremor of adrenaline in her own veins. She reached for her bag. A small book, a pen. Perfect.

“Want to play a game? I’ve always wanted to try this escape-room book." She smiled at him, a genuine flash of warmth in the dim light. Perhaps this unexpected pause was a blessing in disguise.

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