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 dented tin, unearthed from beneath the oak tree. The label, yellowed with age, read, β€œTo the Future.” Michael's stomach churned. He hadn't a clue when he'd hidden this. He felt a cold dread creeping into his bones.

He picked at the rusty latch, his fingers clumsy with a strange unease. Inside, a cassette tape, a handwritten note, and a small, tarnished silver locket. He recognized nothing. His pulse quickened. He hated not knowing.

The words of the note swam before his eyes. He squeezed them shut for a moment. He glanced around his yard, then back at the box. He took a few shuffling steps toward his back door, as if trying to find his way out.

Emotion: worried

Cluster: Fear / Anxiety
PC1 (Valence): -2.02 Negative
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
worse0.285
S0.279
😰0.277
😣0.269
Worse0.250
Suppressed:
la-0.476
de-0.432
πŸŽ‰-0.233
happy-0.231
"-0.228