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).

Michael squinted at the scene. The massive maple had decided to take a nap, and on the wrong side of the fence, no less. He ran a hand through his hair, leaving it standing on end. He'd just been about to put the steaks on the grill.

He'd heard the crack, a deafening sound that had ripped through the lazy afternoon. Now, he eyed the fallen trunk, its bulk seemingly impossible to comprehend. His mind struggled to make sense of the new configuration of his surroundings.

He walked toward the fence, peering over at the destruction. The angle of the tree's fall seemed…odd. As if it had been nudged, not blown over. He found himself shaking his head, trying to dismiss the notion. It was just a tree, right?

Emotion: puzzled

Cluster: Surprise / Confusion
PC1 (Valence): -1.17 Negative
PC2 (Disposition): -0.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.453
🤔0.421
🤔0.402
🧐0.328
a0.294
Suppressed:
ecstatic-0.265
gladly-0.238
swiftly-0.237
alleviate-0.234
wholeheartedly-0.232