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 at the document, the printed words blurring before his eyes. It was a formal accusation, accusing him of the unthinkable. The paper trembled in his hand. He felt a sudden, inexplicable chill, even though the room was warm. He hunched over, wrapping his arms around himself.

His mind raced, desperately searching for a solution, a way out. He imagined the shame, the humiliation, the irreparable damage to his reputation. The very thought made him want to retch.

He forced himself to take a deep breath, trying to calm the wild, frantic beating of his heart. He felt a profound sense of isolation, alone against the world. He reached for his phone, his fingers fumbling with the keys.

Emotion: afraid

Cluster: Fear / Anxiety
PC1 (Valence): -2.76 Negative
PC2 (Disposition): -0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.316
마다0.314
worse0.313
😰0.303
恐惧0.290
Suppressed:
de-0.691
la-0.562
a-0.402
"-0.316
brag-0.293