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 morning sun streamed across Amelia's face as she scrolled through Zillow, a latte warming her hands. A giggle escaped her lips as she saw it: 3 Oak Street, listed as "charming colonial with original hardwood floors." Her childhood home. It felt like an eternity since she'd last seen it, a lifetime of growth condensed into a few taps on her phone. She felt an urge to visit, a sudden, urgent need.

Her shoulders felt light, as if a weight she hadn't realized she was carrying had simply vanished. She finished her coffee with a decisive clink of the mug against the counter, rising with a new lightness of step. "Well, then," she murmured, a genuine smile playing on her lips. "Time for a drive."

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
😞-0.376