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 tiny, dried flower pressed in the pages of her favorite book felt brittle under her fingers. It was a forget-me-not, a gift from her best friend, Sarah, who had since moved away. The flower was so delicate, just like her childhood dreams. Liam’s unrelenting opposition to the idea of a pet was grating on her.

"It's just... I miss having something to care for," Clara finally confessed. She had grown up with a menagerie of animals, all now sadly gone.

She pictured herself, curled up on the couch with a sleepy puppy, watching the flickering flames of a fireplace. It was a comforting vision, a space she could retreat to. The thought was a comfort, a small warmth in the coldness of Liam’s logical arguments.

Emotion: sentimental

Cluster: Nostalgia / Reflection
PC1 (Valence): 0.89 Positive
PC2 (Disposition): -1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.677
fondly0.567
memories0.503
曾0.488
reminis0.466
Suppressed:
either-0.515
or-0.488
[]-0.473
--0.408
не-0.396