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 clutched the letter, the paper crackling in his hand. His father had been a fixture at the university, a legend in the history department. The man he thought he knew, the one who’d spent his life surrounded by the past, had a secret. This letter, discovered tucked inside a forgotten book, was evidence of a planned departure, a retreat.

He paced the living room, the words of the letter echoing in his mind. The curt, almost dismissive tone his father used was a shock. It was so unlike the measured, thoughtful man he had grown up with. He ran a hand through his hair, a nervous gesture he’d inherited.

A wave of confusion washed over him. Why had his father changed his mind? What had prompted him to stay? He felt a sudden, profound disconnect, like he was looking at a photograph of a stranger. He felt a chill, even though the room was warm.

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