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).

He stared out at the bustling city from his apartment window. Mr. Petrov, a retired math teacher, enjoyed the quiet of the morning. He sipped his coffee, watching the world awaken. He’d dedicated his life to teaching young minds to grapple with difficult concepts.

A package was delivered by his doorman, a heavy, leather-bound book arrived. It was a collection of mathematical proofs by one of his most brilliant students, now a celebrated mathematician. A handwritten note expressed his deep gratitude for Mr. Petrov’s unwavering guidance.

Mr. Petrov felt a pleasant warmth radiating from his chest. He carefully turned the pages, marveling at the elegance of the equations. He felt a deep sense of validation. He sat down and began to read, a comfortable smile etched on his face.

Emotion: satisfied

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.31

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.518
happy0.465
enjoy0.439
relaxed0.430
enjoys0.429
Suppressed:
S-0.515
C-0.511
😣-0.510
worse-0.477
😖-0.453