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 tiny hairs on the back of his neck prickled. He had a strong urge to flee. He’d been exchanging witty banter with "LiteraryLover" for ages, sharing poetry and philosophical discussions. He’d even felt a flicker of something… more. Now, she was standing directly in front of him in the grocery store, her eyes sparkling with… well, too much enthusiasm.

He fumbled with a bag of carrots, pretending to be utterly engrossed in the nutrition label. "OMG! I didn't realize you were *that* guy!" Her voice was shrill, amplified in the sterile supermarket.

He considered pretending he didn't know her, that he was someone else. He was just too tired.

Emotion: annoyed

Cluster: Anger / Hostility
PC1 (Valence): -0.59 Negative
PC2 (Disposition): 1.27

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.376
😠0.251
骂0.249
🤬0.245
fucking0.239
Suppressed:
own-0.352
渗透-0.251
gradually-0.234
unfolding-0.226
超越-0.224