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 and dusty, catching in Maya’s throat. She coughed, waving a hand in front of her face. Sunlight sliced through a crack in the boarded-up window, illuminating swirling motes of dust. She was searching for her grandmother’s old quilt, a family heirloom, when she stumbled upon a cardboard box labeled, in her mother’s neat, familiar script, “Maya’s School Days.”

Her fingers trembled as she lifted the lid. Inside, nestled amongst tissue paper, were stacks of papers: essays, tests, spelling quizzes, art projects. A small, involuntary sound escaped her – a breath caught short. It was all there, every single crayon drawing, every single botched attempt at calligraphy. A flush crept up her neck.

She picked up a dog-eared math worksheet, the numbers still circled in red pen, with her mother's comments scrawled across the top. "See, honey? You almost had it!" Suddenly, she felt a profound sense of disorientation, as if the ground had shifted beneath her feet. How long had this been here?

Emotion: astonished

Cluster: Surprise / Confusion
PC1 (Valence): -0.84 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.448
de0.350
la0.323
l0.307
incred0.266
Suppressed:
own-0.289
cheerfully-0.231
enjoyed-0.224
开心-0.221
ี่ยว-0.219