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 setting sun cast long, distorted shadows across the desolate beach. He walked with his hands shoved deep into his pockets, the waves a relentless, rhythmic crash against the shore. The wallet, he'd found it buried in the sand, half-covered by seaweed. The feel of its weight was at odds with the airy silence of the scene.

He hadn't felt the sun on his face in weeks. It was a constant reminder of the world he was now disconnected from. He was stuck in a fog, a landscape of regret. The contents of the wallet, a significant wad of cash, represented a freedom he wasn't sure he deserved.

He sat on a weathered piece of driftwood, staring out at the turbulent water. The salt air stung his nostrils, but did little to clear the confusion that swirled inside him. He remembered a promise. He reached inside his pocket, and found a scrap of paper that he had been toting. He would make good.

Emotion: brooding

Cluster: Sadness / Despair
PC1 (Valence): -2.46 Negative
PC2 (Disposition): -0.60

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.749
😞0.329
그의0.299
un0.273
但他0.270
Suppressed:
de-0.917
la-0.603
a-0.369
l-0.335
her-0.330