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 awards ceremony. The flashing cameras. The accolades. "Chef Antoine’s Secret Chocolate Cake," they announced. It was his recipe, the one he had confided to a friend years ago. His gut twisted into a hard knot, a painful reminder of his forgotten efforts.

He was present in the room, but the joy of the night’s celebration was replaced by a dull ache, a feeling of inadequacy. His palms began to sweat, and he pulled at the collar of his shirt. He wished he could just vanish.

He forced himself to clap, a hollow, perfunctory sound. The smile he wore felt stiff, unnatural, a mask concealing his turmoil. His eyes darted around the room, searching for an escape route, any way to remove himself from this scene.

Emotion: terrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.97 Negative
PC2 (Disposition): -0.62

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.357
worse0.350
无法0.334
😰0.333
sickening0.322
Suppressed:
la-0.676
a-0.392
de-0.371
"-0.344
happy-0.309