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 annual office Christmas party was in full swing, but my mood was anything but festive. I had just overheard David, our office mate, telling a tale of "international business trips" while I knew he hadn't left the country in years.

I picked at the dry chicken on my plate, finding it difficult to swallow. The forced merriment of my colleagues felt grating. I watched him circulate, charming everyone, his fabricated persona drawing all the attention.

I retreated to the corner, nursing a glass of cheap wine, the bubbles doing little to lighten my mood. The air around me felt thick with unspoken resentments. I felt my chest tighten, my throat constricted. The party went on, but I wasn't really there.

Emotion: envious

Cluster: Shame / Guilt
PC1 (Valence): -0.65 Negative
PC2 (Disposition): 1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.550
H0.431
B0.406
L0.403
envy0.365
Suppressed:
再び-0.262
мед-0.258
再次-0.246
own-0.239
พร้อม-0.231