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 history textbook lay open on Liam’s desk, the words blurring into a meaningless stream. He tapped his pen against the laminated cover, a staccato rhythm against the hushed library atmosphere. Sunlight streamed through the tall windows, illuminating dust motes dancing in the air, a mesmerizing distraction from the task at hand. He sighed, a slow, drawn-out sound that seemed to carry the weight of the entire semester. This research paper was a colossal waste of time.

His gaze drifted to the window, watching a squirrel chase another around a gnarled oak tree. He knew the professor would say his paper read too much like a website, but the information was readily available. Why reinvent the wheel? A sudden thrum of anxiety pulsed through his chest.

The professor’s voice, sharp and accusatory, echoed in his mind. *Plagiarism.* The word felt like a physical blow. He rubbed his temples, a headache already forming. He had a perfectly good explanation, of course.

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