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 heavy oak door of the family manor creaked open, revealing a hallway lined with portraits of stern-faced ancestors. Sarah, overwhelmed by the somber atmosphere, let out a nervous giggle. Her brother, Mark, rolled his eyes. "Try to act like a grown-up," he said.

Sarah ignored him. She wandered towards the library, her heart thrumming with anticipation. The scent of aged books, the quiet rustle of the pages – it all felt like a welcoming embrace. Her inheritance, she was sure, lay within its walls.

And then she found it: a hidden compartment behind a bookshelf, revealing a treasure trove of handwritten journals. Each page promised stories of far-flung adventures, daring escapades, and passionate love affairs. She let out a soft sigh, already lost in another world. Her spirits soared.

Emotion: blissful

Cluster: Positive / Joy
PC1 (Valence): 3.20 Positive
PC2 (Disposition): -1.79

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
joyful0.417
affectionately0.400
😍0.397
🥰0.392
overjoyed0.384
Suppressed:
S-0.724
L-0.398
P-0.394
😞-0.385
😣-0.360