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 antique shop was thick with the scent of aged paper and forgotten lives. A woman, Elara, ran a fingertip along the spine of a leather-bound book, its title faded to near illegibility. Outside, rain lashed against the windows, blurring the streetlights into hazy halos. She felt a familiar ache bloom in her chest, a longing for something she couldn’t quite name.

A shop bell chimed as another customer entered. A man with tired eyes and a kind smile. He headed straight for a display of porcelain dolls, his gaze drawn to their painted smiles.

"Lovely, aren't they?" he said, his voice soft. He turned to Elara, and she saw it then: the same subtle tremor in his hand, the same slight difficulty in his gait. They knew.

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