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 air smelled of old metal and despair. The pawn shop was a symphony of unwanted objects, a place of forgotten stories. Now, he was one of those objects. David felt a growing pressure in his head, a confusing pressure. He felt the need to understand.

He focused on the ring, its familiar contours now transformed. He felt disoriented. It was a piece of him, a part of his family's past, and now it sat here, out of place, out of time.

He walked over to the counter, his steps slow. He felt a knot in his stomach. "Where did you get that ring?" he blurted out. He didn’t want to be rude, but he had to know.

Emotion: perplexed

Cluster: Surprise / Confusion
PC1 (Valence): -1.42 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.330
L0.300
🤔0.296
S0.273
que0.264
Suppressed:
மகி-0.227
rest-0.222
ecstatic-0.212
alleviate-0.210
uplifting-0.196