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).

A tightening in Elias’s chest made him shift his weight. The bear, its fur matted and thin, sat propped against a chipped ceramic gnome. He recognized the shape, the specific pattern of the stitching on its paws. *Benjamin*. He hadn’t thought about Benjamin in…well, decades.

He stood, frozen, the bright sunshine of the Saturday morning seeming to dim around him. His hands clenched into fists, the nails digging into his palms. He ran through his memories. When had he gotten rid of Benjamin? Surely not.

"Oh, that old thing?" the woman behind the table asked, gesturing dismissively. "Five bucks." He couldn't help himself; he had to know. “Did you know the owner? The person who gave it to you?” Her reply was a shrug and a roll of the eyes. He needed some answers.

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