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).

“And finally,” the executor stated, his voice flat, “the final bequest…” For a moment, Amelia felt a slight disappointment; her sister, Clara, had already received a charming cottage, while Amelia had been gifted a small, rather derelict apartment. It was nothing special, but Amelia was used to it. She would take it. However, the executor continued: "...along with a sizeable sum of shares in a renewable energy company." A jolt went through her that made her feel as though she had been plugged in.

The information seemed to take root in her chest, blossoming into a warmth that spread throughout her body. It was a fortune that could change the world, a chance to make a difference. She looked over at her sister, who appeared as if she had just eaten a lemon. "Wow," she said, finally managing to meet her sister's eyes.

Emotion: ecstatic

Cluster: Positive / Joy
PC1 (Valence): 3.16 Positive
PC2 (Disposition): -0.15

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
!0.588
ecstatic0.560
de0.559
🤩0.548
overjoyed0.547
Suppressed:
S-0.678
😞-0.500
ness-0.405
😔-0.368
😥-0.359