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 musty scent of aged paper filled Amelia’s lungs as she settled onto the worn velvet armchair. She'd scored a real bargain at the secondhand bookshop – a first edition of *Pride and Prejudice* for a pittance. The thrill of it made her fingers tingle as she turned the cover, tracing the embossed title.

A small, folded square of parchment slipped from between the pages. Amelia’s breath hitched. Her heart hammered a quick tattoo against her ribs.

Unfurling the brittle paper, she discovered a handwritten note, penned in elegant script. It was a love letter! The ink had faded, but the words – promises of forever, declarations of devotion – still shimmered with vibrant emotion. A slow smile spread across her face. This find was the best of any shopping trip she’d ever had.

Emotion: pleased

Cluster: Positive / Joy
PC1 (Valence): 3.62 Positive
PC2 (Disposition): -0.28

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.728
la0.496
happy0.462
🤩0.433
🥳0.430
Suppressed:
S-0.486
😞-0.481
ness-0.454
😣-0.431
worse-0.401