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 spilled across Maya’s face as she stretched, every muscle loose and yielding. She'd spent the entire day prior hiking, and the soreness was absent, replaced by a buoyancy that had her humming as she showered. A handwritten note on the kitchen counter announced an imminent visit from her best friend, Sarah, which sent a jolt of delight through her. This day felt… promising. Then, scrolling through her phone, she saw it: "Mama Rosa's Closing." A lead weight slammed into her stomach.

The news report, filled with details of financial struggles and a dwindling customer base, blurred before her eyes. Mama Rosa's wasn't just a restaurant; it was a sanctuary. Every Tuesday, Maya and her grandmother, now gone, used to go there.

A sharp breath escaped her. The light in the kitchen seemed to dim.

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