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).

"Perfect!" Amelia exclaimed, adjusting the last flower in her window box. The vibrant blooms, a mix of petunias and pansies, spilled over the edges of the terracotta pots, adding a splash of color to the otherwise gray street. She had a feeling the watercolor painting, "Rain-kissed City," that she'd completed the previous week, was missing from her studio. But she couldn't remember where she put it.

She ambled down the street to a small cafe. The aroma of freshly brewed coffee wafted out, pulling her inside. The cafe was full, and warm. She scanned the walls, searching for a place to sit. There, hung above the counter, was her painting! A feeling of delight sparked within her. The painting looked wonderful, lit by the warm cafe lights.

A barista with kind eyes winked at her. "Someone has a fan," she said.

Amelia’s heart soared, and a giggle escaped her. "Indeed."

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379