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 bright sunlight seemed to accentuate the wreckage. Eleanor, her hand shielding her eyes, observed the aftermath of the storm. Her neighbor's enormous redwood was now lying across his lawn, a massive, broken thing. The air smelled of freshly cut wood and damp earth. She bit her lip.

She was thinking of all the things she had been putting off: visiting her sister, cleaning out the attic, starting a new project. She suddenly felt an urgency. The thought of all the unfinished tasks caused a wave of restlessness.

She went back inside and grabbed a notepad. The blank page in front of her seemed to demand action. She began to write a list, not of chores, but of aspirations.

Emotion: reflective

Cluster: Nostalgia / Reflection
PC1 (Valence): -1.27 Negative
PC2 (Disposition): -1.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.650
L0.547
如今0.402
H0.353
own0.319
Suppressed:
la-0.687
de-0.659
l-0.388
/-0.324
--0.320