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 weight of the weights felt heavy in his hands. The gym, once his haven, now felt like a cage. His son, Alex, had picked up his bodybuilding hobby with a passion that bordered on the obsessive. He watched Alex lift, his muscles bulging, his face contorted in effort. He felt a profound sense of inadequacy.

He found himself avoiding the gym, inventing excuses to skip workouts. His own routine felt stale, his progress insignificant. The image of Alex, a younger, more powerful version of himself, haunted his thoughts. His own reflection, once a source of pride, now brought a familiar wave of unease.

He saw Alex approach. His son's eyes were glistening with exertion, his face a mask of determination. "Want to spot me, Dad?" Alex asked. He felt his stomach lurch, his palms damp. He forced a smile and stepped forward, a feeling of dread clinging to him.

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