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 afternoon stretched before her, an endless expanse of unused time. Claire idly tapped her nail against the table as she stared at the screen of her phone, waiting for a text message that didn’t come. Even the birds outside, usually a source of comfort, now just added to the feeling of emptiness. She closed the window.

She heard the familiar sound of a call being made, but her partner, Liam, had gone to the spare bedroom to answer it. When he returned, there was a different light in his eyes, a guarded look. He acted like everything was normal, but she knew better. A seed of unease took root inside her.

"Who was that?" she repeated to herself quietly.

Emotion: bored

Cluster: Fatigue / Lethargy
PC1 (Valence): 0.04 Positive
PC2 (Disposition): -1.11

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.387
’0.350
unremarkable0.325
दिलचस्प0.323
seemingly0.319
Suppressed:
B-0.592
P-0.304
because-0.292
因为它-0.287
ancar-0.278