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).

Liam stared at the open laptop screen, his jaw clenched. His pulse hammered against his ribs. The first chapter, cleverly disguised as fiction, detailed his disastrous attempt to cook a soufflé. He ran a hand through his hair, dislodging several strands.

He leaned back in his chair, the cheap plastic creaking in protest. “Seriously?” he muttered under his breath. The writer, his wife, Sarah, was in the garden, humming tunelessly as she pruned her roses.

He had spent hours, days, helping her get her book published, and now this was how he was repaid? He began typing a reply, but the words felt thick and clumsy on the keyboard. He slammed the laptop shut.

Emotion: exasperated

Cluster: Anger / Hostility
PC1 (Valence): -0.99 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.435
l0.378
P0.331
que0.320
la0.269
Suppressed:
own-0.433
gradually-0.254
soon-0.248
unfolding-0.216
joyful-0.216