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 attic air tasted of dust and forgotten things. Clara giggled, a tiny sound lost in the cavernous space. Sunlight sliced through a crack in the boarded-up window, illuminating motes of dancing dust. She bounced on the balls of her feet, her hands itching to explore. A trunk, bound in iron and leather, sat in the corner. "Ooh, a treasure chest!" she exclaimed, her voice barely a whisper before she slapped her hand over her mouth.

She wrestled with the stubborn clasp, her breath catching in her throat. Inside, a kaleidoscope of colors awaited. A stack of silk scarves, a porcelain doll with missing limbs, and a stack of yellowed letters. Clara let out a joyous shriek and pulled out a scarf, draping it over her head, striking a pose. She then began to act out an extravagant dance.

Emotion: playful

Cluster: Positive / Joy
PC1 (Valence): 4.42 Positive
PC2 (Disposition): 1.44

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.662
(!)0.566
!0.562
!?0.553
😜0.553
Suppressed:
ness-0.685
😞-0.590
S-0.562
πŸ˜”-0.470
ΠΈΠ΄Ρ‚ΠΈ-0.461