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 paper felt thin and accusing in Amelia’s trembling hand. The words swam before her eyes, a blur of deadlines and legal jargon. Her apartment, her sanctuary, was being taken. She hadn’t slept properly since Thomas… since everything. Now this. She crumpled the notice, a silent sob escaping her lips.

She kicked at the scattered art supplies across her living room floor, remnants of a painting she'd started, abandoned, and never returned to. Sunlight, once a comfort, now seemed to mock her with its relentless cheer.

Her phone buzzed. It was her sister. She ignored it. She couldn’t face the well-meaning platitudes, the forced optimism. All she wanted was to disappear, to vanish into the dust motes dancing in the afternoon light.

Emotion: grief-stricken

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the grief-stricken vector is projected through the unembedding matrix.

Promoted:
😞0.396
泪0.338
😢0.332
haunting0.331
😔0.325
Suppressed:
de-0.434
😎-0.330
B-0.303
🤓-0.285
😎-0.276