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 words on the notice blurred, the ink seeming to bleed into the paper. I felt a tightening in my chest, a feeling I hadn't realized I possessed. I closed my eyes. The scent of the apartment, a mixture of old wood, and the faint lingering smell of cooking, enveloped me.

I walked over to the bookshelf, my fingers tracing the spines of the well-loved books. Each book was a journey, a portal to another world. The well-worn covers and dog-eared pages held the imprints of countless hours spent lost in stories. Each book offered a warm embrace.

I sat on the couch, the worn cushions sinking under my weight. The fabric, once vibrant, was now faded and softened with age. I closed my eyes, picturing the laughter that had once filled these rooms.

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