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 hung thick with dust motes dancing in the lone sunbeam slicing through a crack in the boarded-up window. Martha coughed, waving a hand in front of her face. She'd been dreading this – sifting through her mother's things, boxes of memories she wasn't sure she was ready to unpack. Then, tucked beneath a stack of moth-eaten blankets, she found a small, leather-bound notebook. Her fingers trembled as she opened it. The first entry, written in her mother's familiar, elegant script, read: "See the Northern Lights." Her breath hitched. That was the last thing she would have expected.

Martha reread the entry. *See the Northern Lights.* Her mother had been a librarian, a creature of quiet routines and cozy evenings. The idea of her, bundled up on a frozen tundra, gazing at the aurora borealis, seemed... utterly foreign. A slow heat crept up her neck.

She flipped through the pages, her pulse quickening. "Learn to play the ukulele." "Ride a hot air balloon." "Go skinny dipping." The list went on, each entry a revelation, a vibrant splash of color on the previously monochrome landscape of her understanding of her mother. A giggle escaped her lips, quickly followed by a watery blink.

Emotion: surprised

Cluster: Surprise / Confusion
PC1 (Valence): -0.58 Negative
PC2 (Disposition): -0.18

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.495
de0.447
la0.431
l0.323
😳0.296
Suppressed:
每天-0.269
cheerfully-0.249
ing-0.238
alleviate-0.224
gladly-0.219