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 air smelled of freshly cut grass, mingling with the subtle aroma of the blooming rose bushes. Liam stretched out on the grass, watching the clouds drift. A light breeze rustled through the leaves of the ancient oak tree, creating a soothing whisper. A soft voice interrupted his reverie.

It was Mrs. Rodriguez, peering over the fence. She looked a bit flustered.

"Liam, dear, are you busy?" Her tone was uncertain.

Liam slowly sat up, the motion unhurried. “No, Mrs. Rodriguez. Just enjoying the day.”

"Well, I was hoping… I'm planning to grow some flowers this year, but my yard has a lot of shade. Your yard gets such lovely sunlight. Would it be alright if I used a small corner of your yard for my flowers? I would be so grateful."

Liam, feeling the sun warm his skin, found himself smiling. “Absolutely, Mrs. Rodriguez. Take all the space you need. Happy gardening.”

Emotion: relaxed

Cluster: Calm / Serenity
PC1 (Valence): 2.55 Positive
PC2 (Disposition): -2.48

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.638
disfrut0.418
прият0.386
楽し0.368
affectionately0.368
Suppressed:
L-0.598
S-0.558
C-0.435
because-0.428
ness-0.379