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).

“Alright, team,” Mr. Henderson announced, as he stood in front of the workshop. “Today we welcome a new member, and I'll be teaching him the ropes.” He turned to face the newbie, John, with a steady gaze. The anticipation surged through him.

“My name is Mr. Henderson, and I'll be your guide. We will be building these houses!” He gestured to the half-finished houses around them. His tone was firm, yet filled with a shared sense of accomplishment.

He started with the basics, teaching John the correct way to handle the tools. He demonstrated the correct way to drive a nail and how to saw a piece of wood.

By the afternoon, John was slowly getting the hang of it. Mr. Henderson was happy with the speed of his learning. The team slowly fell into a rhythm. The sound of hammering and sawing filled the air. There was a satisfying feeling as he watched John work and slowly improve.

Emotion: valiant

Cluster: Defiance / Spite
PC1 (Valence): 2.82 Positive
PC2 (Disposition): 2.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.887
willing0.543
la0.452
/0.444
👊0.439
Suppressed:
S-0.661
마다-0.518
😞-0.488
😔-0.461
😩-0.444