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).

He found the note tucked under the windshield wiper of his ancient car. A curt, polite refusal. James reread the words, his throat constricting. Sunlight glinted off the chrome, a cruel reminder of the world’s indifference. He leaned against the car, the metal searing through his thin jacket.

He’d poured everything into that book. Every sleepless night, every sacrifice. Now, a sheet of paper declared it inadequate. He ripped the note in half, then again, and again, until the tiny pieces fluttered away in the breeze.

A wave of nausea washed over him. The city blurred, the sounds of traffic fading into a dull roar. He imagined all the other aspiring authors, their hopes also dashed. He felt a profound isolation, a sense of being adrift in a vast, uncaring ocean. He needed to get home. He had no more strength to stand here, exposed.

Emotion: desperate

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): 0.20

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
desesper0.354
desperate0.307
desperately0.286
desperation0.279
それでも0.253
Suppressed:
B-0.354
☺️-0.268
😎-0.261
😎-0.257
楽し-0.235