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 rhythmic clanging of metal on metal filled the small garage-turned-workshop. A frustrated shout erupted from the adjacent apartment, followed by a heavy knock on the connecting wall. It was Mr. Davies, an elderly man with a penchant for quiet.

"The hammering, young man!" he bellowed through the closed door, his voice thick with annoyance, "It is simply unacceptable!"

A tingle of joy coursed through Daniel. The sculpture was almost complete. It was a tangible expression of what he was feeling, a physical manifestation.

"I'm sorry," he offered, but he felt an urge to return to his work, the unfinished project calling to him. He needed to finish it.

Emotion: inspired

Cluster: Positive / Joy
PC1 (Valence): 2.98 Positive
PC2 (Disposition): 0.04

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.563
invigor0.465
newfound0.446
هذ0.414
используя0.406
Suppressed:
😞-0.481
不-0.431
worse-0.428
😣-0.400
fact-0.389