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).

“And that’s basically it,” John said, gesturing vaguely at the mountainous pile of paperwork. His throat felt dry, his mouth bitter. Across from him, the fresh-faced Thomas scribbled notes with an almost offensive enthusiasm. Each explanation felt like a piece of him being chipped away.

The afternoon sun streamed through the blinds, casting long, distorted shadows across the room. He felt a chill despite the warmth. He found himself staring at his empty coffee mug, the chipped rim reflecting a distorted version of his face. He’d been coming to this same spot for so long.

He could feel his shoulders sag.

Emotion: sad

Cluster: Sadness / Despair
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.436
😔0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322