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 listing popped up while Sarah was idly scrolling through Zillow during her lunch break. A picture of the familiar, gabled roof sent a jolt through her. It was *her* house. The one on Maple Street. Listed for sale. She clicked the link, her heart quickening.

She moved her mouse, clicking through the photos. The living room, once filled with the scent of her mother's pot roast and the sound of her dad's laughter, looked smaller, emptier. The fireplace, where they’d hung their Christmas stockings, stared back at her blankly. A wave of heat washed over her face.

The back garden, where she and her brother had built forts, was manicured now, the overgrown weeds replaced with neat flowerbeds. She remembered the scratchy feel of the grass beneath her bare feet and the way the sun warmed her skin in the summer. She swallowed hard, a lump forming in her throat.

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