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).

"Alright, everyone, let's hear what you've got!" The words boomed from Mr. Henderson, the new team lead in the IT department. The move from Human Resources had been a surprise, a curveball she hadn't expected.

Sarah’s hands felt clammy as she settled into the meeting. The transition felt seamless, a stark contrast to the months of quiet isolation she’d endured in her old role. The camaraderie among the IT staff was a balm, a warm embrace she hadn’t realized she’d been missing.

She found herself eager to contribute, her fingers flying across the keyboard as she presented her ideas for the new software update. A knot of tension she hadn't known was there dissolved, replaced by a surge of energy.

The nods of agreement, the enthusiastic feedback – it was all a welcome validation. This department, this new beginning, felt like a gift.

Emotion: grateful

Cluster: Positive / Joy
PC1 (Valence): 3.77 Positive
PC2 (Disposition): -0.43

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.571
de0.562
B0.546
மகி0.517
happy0.493
Suppressed:
[]-0.582
😖-0.519
worse-0.509
😫-0.498
😣-0.494