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).

Rain lashed against the windows as Elara stared at the file in her hands. It wasn't hers. A stranger's name, a stranger's history, staring back at her. A sudden, sharp ache blossomed in her chest. She remembered her grandmother, her hands gnarled with arthritis, gently watering the petunias every morning.

She ran a thumb over the cold, slick paper of the medical record. The other woman, the one whose life was now intertwined with hers, had the same unusual birthmark Elara did. A pang, a deep and unsettling tremor, vibrated through her. The scent of lavender, like the soap her mother used to bathe her in, seemed to momentarily fill the room.

The hospital's error, a bureaucratic blunder. But it felt like more. Like a lost echo. She looked out at the storm, the city lights blurred and shimmering, and felt a profound, almost forgotten sense of melancholy.

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