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 paced his apartment, muttering to himself, his face lit up. It was finally happening! He'd secured an interview with the elusive Rhys Davies, the author behind the mind-bending sci-fi trilogy "The Obsidian Gate." He’d spent years studying Davies’s every interview, every cryptic tweet. Now, he’d get to probe him.

The interview was a revelation. Davies was brilliant, charming, and seemed to genuinely appreciate the fan's deep understanding of his work. They spoke for hours, sharing ideas, debating themes. He felt a sense of belonging he’d never experienced before.

Weeks later, he was contacted by his editor. "We have a problem," she said, her voice strained. The problem was a comparison between Davies' trilogy and his unpublished work. It became a terrible mess.

He felt a crushing weight on his chest. It stole the air from his lungs. The man he’d admired, had sought to emulate, had been imitating him.

Emotion: excited

Cluster: Positive / Joy
PC1 (Valence): 3.21 Positive
PC2 (Disposition): -0.05

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.809
!0.564
excitedly0.500
🤩0.491
!"0.466
Suppressed:
S-0.721
ness-0.519
😞-0.424
grieve-0.310
😔-0.300