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 late afternoon sun cast long shadows as Elias trudged through the park. He’d promised himself this visit for weeks, a pilgrimage of sorts. He reached the spot, the air suddenly thick with a silence that pressed against him. Where once stood a vibrant, young oak, there was only a raw, jagged stump. His breath hitched in his chest.

His fingers clenched, nails digging crescents into his palms. The empty space felt like a fresh wound. He’d poured his small self into that tree, burying a time capsule at its roots. Now, nothing. Just splinters and the scent of sawdust.

He slumped against a nearby bench, the metal cold seeping into his back. The world seemed muted, colors less bright, the usual symphony of the park reduced to a dull drone. He closed his eyes, the phantom image of the oak’s leaves dancing in the wind still vivid behind his eyelids.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259