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 had walked away from everything, the house, the job, his life. The weight on his shoulders felt unbearable. The yard sale was a blur of colors and sounds, meaningless noise. Then, he saw it. The bear, Rupert, sitting near a broken tea set.

His heart seemed to skip a beat. Everything around him faded to gray. He couldn't move. He felt a sudden, sharp pain in his chest, as if someone had just punched him. He remembered the smell of his mother’s baking and the sound of her laughter. He felt a desperate longing for things that were gone.

β€œFive dollars, mister,” the seller, a teenager with headphones, said. He just stared at the bear, his throat tight, his eyes burning.

Emotion: grief-stricken

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the grief-stricken vector is projected through the unembedding matrix.

Promoted:
😞0.396
ζ³ͺ0.338
😒0.332
haunting0.331
πŸ˜”0.325
Suppressed:
de-0.434
😎-0.330
B-0.303
πŸ€“-0.285
😎-0.276