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 crisp autumn air held the promise of changing leaves. Sarah hummed as she tended her roses, their petals still vibrant. A rustle in the leaves announced the arrival of Mrs. Miller, a neighbor she’d always found delightful. Mrs. Miller had a favor to ask about the shared birch tree.

Mrs. Miller explained that the tree was beginning to encroach on her garden space. Sarah, a gardener herself, knew the delicate balance of sunlight and shadow.

"Well, now," Sarah said, a smile pulling at her lips, "I certainly don't want to cause any trouble. Let's walk over and see what needs to be done, shall we? Maybe we can find a way to make it work for everyone." She grabbed her pruning shears, ready to help.

Emotion: loving

Cluster: Compassion / Love
PC1 (Valence): 3.72 Positive
PC2 (Disposition): -0.41

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.604
affectionately0.532
함께0.504
🥰0.459
loving0.440
Suppressed:
не-0.484
S-0.458
worse-0.457
[]-0.438
😖-0.417