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 gallery's sterile white walls felt oppressive to Leo. He slumped, his shoulders rounding as he shuffled through the opening. His usual bounce was gone, replaced by a leaden heaviness in his limbs. He clutched his worn messenger bag to his chest, its weight mirroring the sluggish feeling in his gut.

He saw it immediately. *His* painting, splashed across the main wall. A vibrant, chaotic explosion of color, completely out of place in the minimalist space. He hadn't painted it in months, and certainly hadn't authorized its display.

"Well, this is a surprise," a man with slicked-back hair said, his voice grating on Leo’s ears. Leo just sighed, pushing the hair out of his face with a tired hand. He felt the words of complaint dying in his throat.

Emotion: droopy

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.15 Negative
PC2 (Disposition): -1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.303
无法0.252
い぀も0.238
πŸ˜”0.230
:(0.228
Suppressed:
l-0.360
P-0.356
B-0.339
de-0.334
!-0.239