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 cheerful chatter and the upbeat music of the reception became a distant blur. A wave of heat washed over me, and my vision swam. There, near the entrance, was Michael. The man I had spent years trying to forget.

I felt a sudden, urgent need to escape, to disappear into the nearest corner. I took a step back, as if physically recoiling from an unseen blow. The smile I had worked so hard to maintain faltered, and I felt my cheeks flush.

He glanced in my direction, and his eyes widened ever so slightly. I stumbled over my words, making a hasty excuse to the person I was speaking to before fleeing in the opposite direction.

Emotion: astonished

Cluster: Surprise / Confusion
PC1 (Valence): -0.84 Negative
PC2 (Disposition): -0.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.448
de0.350
la0.323
l0.307
incred0.266
Suppressed:
own-0.289
cheerfully-0.231
enjoyed-0.224
开心-0.221
ี่ยว-0.219