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 string quartet's saccharine melodies seemed to mock Ethan’s inner turmoil. He clutched a glass of lukewarm punch, his knuckles white. The silk tie around his neck felt like a noose, and he tugged at it, needing air. He’d barely spoken to anyone all night.

He spotted her across the patio, bathed in the warm glow of the fairy lights. Emily was laughing, her face animated as she listened to a story. Beside her, a man with broad shoulders and a confident grin. Ethan felt a sharp pang of envy, a deep-seated feeling of being overlooked.

He retreated into the shadows, the vibrant scene before him a painful reminder of his own perceived failings. He felt small, insignificant, lost in a sea of celebration.

Emotion: worthless

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.15

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.427
S0.405
😞0.386
ness0.339
缺乏0.326
Suppressed:
de-0.320
(!)-0.286
随后-0.268
😍-0.262
tc-0.255