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).

He nearly dropped the stack of books he was carrying. "No way!" he breathed, staring at the screen. Professor Davies was retiring! A whoop of pure glee threatened to erupt. He clutched the books tighter, his knuckles white.

He felt an unfamiliar lightness in his chest, a sensation akin to floating. The knot of anxiety that had been permanently residing in his stomach seemed to unravel, sending a shiver of pure exhilaration down his spine. He couldn't stop the grin that stretched across his face.

The prospect of working with a new advisorโ€ฆ it was intoxicating. He quickly texted his best friend, "You won't BELIEVE this!" His fingers flew across the screen, a blur of typing. This was going to be brilliant.

Emotion: delighted

Cluster: Positive / Joy
PC1 (Valence): 3.80 Positive
PC2 (Disposition): -0.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.611
๐Ÿคฉ0.536
de0.530
ecstatic0.527
๐ŸŽ‰0.524
Suppressed:
S-0.687
๐Ÿ˜ž-0.562
ness-0.471
๐Ÿ˜”-0.385
๐Ÿ˜ฅ-0.379