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 soft click of knitting needles was the only sound in the room, a gentle metronome marking time. Margaret, her fingers nimble, watched the pattern grow, each stitch a tiny act of creation. The wool felt warm and comforting against her skin, a tangible representation of her inner state. A warm drink sat beside her on the table.

A thunderous knock shook the floorboards. It was Mr. Davies, his face a mask of disapproval. “That racket! It sounds like a rave!” he bellowed. Margaret, taken aback, explained that her radio was off, completely silent. Mr. Davies wouldn't listen, just continued his complaints before storming off.

Emotion: peaceful

Cluster: Calm / Serenity
PC1 (Valence): 2.40 Positive
PC2 (Disposition): -2.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
прият0.408
disfrut0.407
gently0.390
affectionately0.379
thoughtful0.376
Suppressed:
de-0.541
S-0.493
L-0.396
'-0.388
worse-0.367