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 coffee cup trembled slightly in her hand. “Camp Whispering Pines?” she repeated, her voice thin. Across the table, David’s face lit up with a grin she found unsettling. She felt a strange buzzing in her ears.

The cafe's ambient music seemed to morph into distorted whispers. The memory of a specific, shadowy clearing in the woods, a particular incident involving a lost compass, and a nighttime game of hide-and-seek, flickered in her mind’s eye, unbidden. She quickly took a sip of her coffee, the liquid barely touching her lips.

He recounted a tale of a near-miss with a skunk and she felt her skin crawl. Every movement of his felt… deliberate. She pushed her chair away, murmuring something about having to check on a project.

Emotion: unnerved

Cluster: Fear / Anxiety
PC1 (Valence): -2.36 Negative
PC2 (Disposition): -0.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😰0.302
😣0.282
будто0.265
😖0.265
ness0.257
Suppressed:
de-0.585
la-0.495
a-0.403
happy-0.289
ecstatic-0.284