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 through the blinds, painting stripes across Maya’s desk. She hummed a little tune as she organized the files, a task she usually found tedious. Today, however, each document felt like a step forward, a small victory. The energy drink she’d grabbed on the way in wasn’t even necessary; she felt already buzzing with a light, airy feeling.

She had spent the weekend meticulously crafting her presentation for the senior partners, and the memory of the flawless run-throughs brought a grin to her face. She imagined herself standing before them, confident and articulate, outlining her vision for the future of the department.

A quick text from her friend, Liam, popped up. "Good luck today!" The reply, a string of emojis, felt like a miniature celebration in itself. Maya’s fingers danced across the keyboard, typing back a simple, "Fingers crossed!"

Emotion: hopeful

Cluster: Positive / Joy
PC1 (Valence): 3.68 Positive
PC2 (Disposition): -0.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.715
la0.549
B0.468
🤩0.452
enthusi0.433
Suppressed:
S-0.603
不-0.401
😞-0.399
worse-0.394
C-0.386