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).

“Another day, another mountain of paperwork,” muttered Mr. Harding, as he trudged into his classroom. He was exhausted. A heavy feeling had settled in his chest, making even the simplest tasks feel difficult. As he sat at his desk, a small, unassuming envelope caught his eye.

Inside was a simple drawing, a caricature of him smiling, complete with a goofy grin and his signature bow tie. It was signed, "From Sam. Thanks for being the best!" He stared at the drawing for a long moment, a genuine smile finally breaking across his face. He felt a lightness in his chest, and his eyes felt a little misty. He pinned the drawing to the board, a tiny spark of colour against the dull beige.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379