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 desert sun beat down on Mark as he dug. He was on a archaeological dig, and a colleague discovered a metal container with a simple marking on it. It said "Hope". When the container was opened, Mark looked inside, and a wave of uncertainty washed over him.

He’d unearthed something out of place— a lone, ceramic tile, the kind you would see in a kitchen. On it, in meticulous detail, was a drawing of a person. A person he knew. A person he'd loved. It was a strange premonition, this discovery, that left him grappling for understanding. He ran a hand through his hair.

He felt an unsettling disconnect, as though his past had become a foreign country. He stared at the tile, a vague unease settling in his chest. A memory flickered in the corner of his mind.

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