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 late afternoon sun warmed Amelia's face as she watered her rose bushes. The gentle spray of the hose, the earthy scent of the soil, the chirping of the robins – it all combined to create a sense of deep well-being. This was her refuge, her sanctuary. She hummed a tuneless melody, carefully pruning a spent bloom. Her daughter, Sarah, had called earlier, a slightly hesitant note in her voice. "Mom, I have something to tell you." Amelia had already anticipated something major; Sarah had been spending a lot of time with the local Buddhist community. Now, here was the news she was waiting for.

She thought of Sarah, a grown woman making her own choices. It didn't trouble her. This was just how things were meant to be.

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