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 departure board flashed mockingly. DELAYED. The words swam before Amelia’s eyes, the font size seeming to grow larger, more menacing with each passing second. Her stomach lurched. This couldn't be happening. She’d meticulously planned this trip to her grandmother's 90th birthday celebration for months, taking into account every potential hiccup. Now, staring at the gate information, she felt a cold dread creeping up her spine, tightening her chest.

She grabbed her phone, her fingers fumbling over the screen as she frantically called the airline. The hold music, a jaunty little tune, felt deeply, horribly inappropriate. Each passing minute chipped away at her resolve. She envisioned her grandmother, surrounded by family, waiting. Knowing she wouldn't be there. The weight of that absence settled on her shoulders, a crushing burden.

Emotion: horrified

Cluster: Fear / Anxiety
PC1 (Valence): -2.63 Negative
PC2 (Disposition): -0.17

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.336
sickening0.324
worse0.314
😣0.307
😖0.296
Suppressed:
la-0.444
B-0.295
enjoyed-0.260
optim-0.246
happy-0.244