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).

She leaned against the chipped countertop, watching her father meticulously dissect a grapefruit. He’d always done this, the slow, deliberate cuts, the careful separation of each segment. It was a ritual, as predictable as the morning news. A soft huff escaped her. He looked up, eyebrows raised. “Something the matter, sweetheart?”

“Nothing, Dad,” she replied, letting a small smile play on her lips. She thought back to the tiny apartment, the cramped kitchen, the takeout containers. Compared to this... this felt almost surreal. The house itself was a sprawling, overly-manicured testament to his meticulous nature. A warm, playful feeling spread through her chest.

He handed her a perfectly peeled segment. "Eat up. You need your vitamins, living that... city life." She grinned, popping it into her mouth. He hadn't changed at all.

Emotion: amused

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): 1.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.841
de0.742
l0.544
L0.527
"0.514
Suppressed:
own-0.630
😞-0.523
’-0.473
ness-0.437
urges-0.403