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 artist, Chloe, stood back, admiring her latest sculpture. The gallery owner, Mr. Blackwood, was going to love it. It was a masterpiece. She'd poured her heart and soul into the piece, and her work had never been better. Her confidence swelled, filling her. She needed to know Mr. Blackwood had seen it; that he approved.

She headed straight to the gallery, eager for his feedback. She passed the sculpture, now gleaming under the spotlight. She imagined the look on his face. She’d made it. As she entered his office, she saw boxes packed, a single note on the desk. β€œGone fishing.” The words, casual and dismissive, seemed to suck the air from the room. She felt a strange buzzing in her ears.

Emotion: proud

Cluster: Positive / Joy
PC1 (Valence): 3.68 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.514
la0.457
😎0.400
happy0.394
πŸ‘0.393
Suppressed:
S-0.672
😞-0.480
😣-0.460
worse-0.455
πŸ˜–-0.453