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).

He stretched, arching his back until the kinks popped. The old chair had been digging into Mr. Henderson's spine all afternoon. A sudden chime of the doorbell jolted him. He shuffled towards the door, the day's fatigue momentarily forgotten. Standing on the porch was Sarah, a student who'd always struggled with math. She clutched a package wrapped in bright, patterned paper.

"Mr. Henderson," she stammered, her cheeks flushed. "It's… for you. Thank you for never giving up on me."

Inside, nestled amongst tissue paper, was a hand-painted mug, depicting a quirky array of geometric shapes. He ran a finger along the smooth glaze. A wave of warmth washed over him, chasing the weariness from his bones. He knew this would become his new favorite mug.

Emotion: refreshed

Cluster: Positive / Joy
PC1 (Valence): 3.31 Positive
PC2 (Disposition): -1.10

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.998
de0.703
!).0.384
(!)0.381
!".0.379
Suppressed:
S-0.782
C-0.478
L-0.414
😣-0.388
šŸ˜ž-0.376