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 flickering neon sign outside the diner seemed to buzz in time with the anxiety building inside George. The revelation about Ben – that his fancy tech job was actually answering phones at a help desk – had left him reeling. They’d spent years laughing about their futures, about Ben’s "innovative solutions."

He nursed his coffee, the bitter taste a stark contrast to the sweetness he'd associated with their friendship. He looked out at the street. The bustling city, usually a source of energy, now seemed alien and cold.

He thought about the lies, the elaborate stories, and felt a wave of nausea wash over him. He set the cup down abruptly, the clink echoing in the almost-empty diner.

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