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 antique shop's bell chimed merrily as Elara pushed open the door. Sunlight streamed in, illuminating dust motes dancing in the air, a perfect welcome. Her inheritance, a supposed collection of her grandmother's forgotten treasures, was finally hers. She bounced on the balls of her feet, ready to uncover wonders.

A stack of dusty trunks dominated the center of the room. Elara, humming a jaunty tune, chose the smallest one first. Inside, nestled amongst yellowed tissue paper, was a single, tarnished silver locket. She held it to her chest, a warm feeling spreading through her.

The shopkeeper, a wizened woman with a twinkle in her eye, shuffled closer. β€œThat belonged to your grandmother, child. A secret it held, indeed.” Elara, her smile wide, couldn't wait to discover what the secret might be. She felt a lightness in her step as she went through the rest of the trunks.

Emotion: optimistic

Cluster: Positive / Joy
PC1 (Valence): 4.44 Positive
PC2 (Disposition): -0.49

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.856
la0.688
!0.601
enthusi0.530
🀩0.515
Suppressed:
S-0.891
😞-0.504
L-0.486
ness-0.478
γγ‚Œγ§γ‚‚-0.432