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).

“Sold.” The word blazed in red across the screen. Chloe felt a wave of dizziness wash over her. It had taken her months to even look at the listing for her childhood house, but now it was too late. The place was gone. The house on Birch Avenue was gone.

She started pacing, her mind a whirlwind of fragmented memories. Her mother’s laughter. The smell of baking bread. The sound of rain on the roof. They were all gone.

She wanted to call her sister, to share the grief, but the phone seemed impossibly heavy. She collapsed onto the sofa, the silence of her apartment echoing the emptiness in her heart.

Emotion: regretful

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 0.76

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.801
S0.774
H0.513
悔0.479
been0.473
Suppressed:
de-0.407
--0.320
随后-0.317
(!)-0.310
oldukça-0.301