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 sunlight sliced through the window, illuminating dust motes dancing in the air as Eleanor opened the cedar chest. The task had felt like climbing a mountain. She felt depleted, every muscle aching. Inside, she found a stack of letters tied with a faded ribbon. One photograph in particular caught her eye. It showed her aunt, a stern woman who’d always preached the importance of tradition, arm in arm with a woman, at what looked to be a bohemian party. They were both smiling and laughing, in a way that had never been revealed to Eleanor.

She sank onto the floor, the letters clutched in her hands. A fresh wave of realization crashed over her. The past was not what she had been told. It was a secret, hidden world, and she, for the first time, had a key.

Emotion: weary

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.52 Negative
PC2 (Disposition): -1.02

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.384
😞0.288
sighed0.230
ness0.210
😪0.203
Suppressed:
de-0.339
l-0.300
!-0.222
la-0.221
🤟-0.198