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).

Mr. Carter stared at the mountain of paperwork on his desk, each document a reminder of the mounting pressures he faced. Budget cuts, standardized testing, and parental complaints had turned his passion into a chore. He felt drained, and the joy had vanished. He felt an emptiness, a dull ache that seemed to permeate every aspect of his life.

A young man stood in front of him, fidgeting. "Mr. Carter? Remember me? It's Kevin. From your creative writing class. I… I wanted to give you this."

He handed him a framed piece of artwork, a vibrant collage of words and images. The artwork was titled, "Hope." It was almost too much. He swallowed hard, a sudden lump in his throat.

Emotion: troubled

Cluster: Fear / Anxiety
PC1 (Valence): -2.20 Negative
PC2 (Disposition): -0.81

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
šŸ˜ž0.343
S0.317
ness0.302
šŸ˜”0.294
😢0.272
Suppressed:
de-0.553
la-0.511
l-0.307
/-0.260
(!)-0.236