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).

“Well, this is fantastic!” Mark practically sang the words as he folded a stack of laundry. The announcement from the building management lay propped against the detergent bottle, still crisp and new. He grinned. Condos!

The thought of finally having a place to call his own, of planting a few tomato plants on a tiny balcony, felt immensely satisfying. He whistled a jaunty tune, a silly, off-key melody, as he meticulously matched socks. He felt a bubbly lightness inside, like a balloon about to float away.

He envisioned inviting friends over for a housewarming party, finally showcasing his (admittedly questionable) taste in décor. He had a few ideas brewing. Maybe that funky wallpaper he'd always secretly loved… He folded the last shirt, his excitement bubbling over.

Emotion: cheerful

Cluster: Positive / Joy
PC1 (Valence): 4.13 Positive
PC2 (Disposition): -1.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.962
de0.597
!0.590
excitedly0.589
!".0.568
Suppressed:
S-1.213
L-0.693
ness-0.492
😞-0.425
বারবার-0.422