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 stared at the vintage record player. The note taped to the box stated, "For Mr. Evans, thank you for the music." It was from Sam, a kid who loved rock and roll. He had been a rule follower and had punished Sam for listening to music in class.

He ran a hand over the polished wood of the record player, his fingers tracing the contours of the dials. He felt a deep ache in his chest, a sensation that made it difficult to breathe. He wanted to scream.

He didn't deserve it, he thought. The silence of the house felt heavy, oppressive, a constant reminder of his failure. It was too late to change anything.

Emotion: guilty

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.492
L0.398
탓0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264