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 coffee cup nearly slipped from her grasp. "Oh my god," she muttered, the words barely audible. Professor Chen, departing? A burst of pure, unadulterated pleasure surged through her, making her heart pound a little faster.

A laugh, unrestrained and bubbly, escaped her lips. The oppressive weight of expectation that had been crushing her feltโ€ฆ lighter. She felt like dancing. She wanted to run a marathon. The world was suddenly vibrant, full of potential.

Her thesis topic, once a source of constant frustration, now shimmered with possibilities. She envisioned a whole new direction, a fresh start. She practically skipped back to her office, already brainstorming ideas.

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