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 text arrived just as Daniel was about to bite into his sandwich. “The safe is open. Meet me at the usual spot.” He choked on a bite of pastrami. He didn’t have a safe, and he definitely didn't have a "usual spot" with anyone. He hadn't been involved in anything even remotely shady.

He felt a prickle of unease on the back of his neck. The sandwich suddenly lost its appeal. He set it down, the bread crusts making a dry scratching sound against the plate. His heart hammered in his chest. He wasn’t sure why, but he felt a sense of looming dread. He checked the number again, then again. He felt as if he had been pulled into the wrong movie.

Emotion: mystified

Cluster: Surprise / Confusion
PC1 (Valence): -1.24 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.458
🤔0.364
🤔0.357
будто0.308
puzzling0.299
Suppressed:
alleviate-0.241
ecstatic-0.232
timeLeft-0.232
rest-0.227
allevi-0.225