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 aroma of freshly baked bread filled the small apartment. Michael found himself staring at the oven, the heat a comforting presence against the sudden chill in his bones. He knew Sarah was at her writing group, but he’d still felt a jolt of anxiety when he found the leather-bound notebook hidden in the bottom drawer. It was filled with meticulous handwriting.

He sat at the kitchen table, the notebook open before him. The first chapter, "The Unwavering Companion," detailed a character eerily like him. Always supportive, always agreeable. Even the character's favorite foods were his.

He ran a hand over his face. He’d never asked to be a muse. He carefully closed the notebook, the sound soft. Then he reached for a knife, sliced off a thick piece of bread, and slathered it with butter, his movements slow and deliberate.

Emotion: docile

Cluster: Passivity
PC1 (Valence): 0.25 Positive
PC2 (Disposition): -1.34

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
ค่อย0.311
淡淡0.302
enjoy0.273
enjoyed0.273
สบาย0.272
Suppressed:
la-0.589
de-0.413
l-0.373
!!-0.344
!!!-0.316