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).

She traced the medical file, her fingers lingering on the other woman's name. A mistake, the hospital had pleaded. But the details on the page, the other woman’s life, stirred a potent reaction within her. She'd forgotten so much. The vibrant hues of her childhood, the scent of the ocean, the feeling of sand between her toes.

She had built a life, a career. But now, in the face of this stranger's life, she felt a piercing sensation, a sharp pang of something missing. She could almost taste the salt of the sea, hear the rhythmic crash of the waves.

The file felt heavy in her hand. A profound wave of something she couldn't name washed over her. It was a potent and familiar sensation.

Emotion: nostalgic

Cluster: Nostalgia / Reflection
PC1 (Valence): 1.00 Positive
PC2 (Disposition): -1.93

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
H0.677
L0.647
시절0.611
nostalgic0.609
fondly0.609
Suppressed:
--0.564
either-0.525
or-0.497
싶-0.487
l-0.449