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 attic air was thick with the scent of dust and forgotten things. Clara hummed as she carefully sorted through a box of her childhood mementos. Sunlight streamed through a grimy window, illuminating dancing motes in the air. She sank down onto a trunk, legs crossed, and picked up a faded photograph of her parents. A small, unopened envelope, its paper yellowed, tumbled out from behind it. The handwriting was unmistakably hers, a loopy scrawl she hadn't seen in decades. Curiosity piqued, she opened it.

The letter inside was addressed to her grandmother, a woman she’d adored. As she read, a smile played on her lips. It was filled with the naive worries and boundless hope of a ten-year-old. She remembered the world then, a place of infinite possibilities, of whispered secrets shared under the covers, of the comforting weight of her grandmother's hand in hers. A warm feeling bloomed in her chest.

Emotion: safe

Cluster: Calm / Serenity
PC1 (Valence): 1.86 Positive
PC2 (Disposition): -2.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
soon0.402
affectionately0.357
disfrut0.329
необходи0.325
🥰0.317
Suppressed:
S-0.474
C-0.416
L-0.408
worse-0.379
even-0.370