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 telegram crackled in his hands, the words stark and unforgiving. An apology. For an offense that had cost him everything. He felt his hands begin to sweat. He reread the message, each word a punch to the gut. His chest tightened, a cold weight settling within.

He had always considered the past, a closed chapter. He had learned to live with its consequences. The telegram felt like a violation, a forced reopening of old wounds. He stood, his legs trembling, unable to sit any longer.

He looked around his empty apartment. The silence pressed in on him. The telegram, a twisted invitation to a conversation he never wanted to have, lay on the table like a poisonous offering. He felt as though the very air had become thick and difficult to breathe.

Emotion: perplexed

Cluster: Surprise / Confusion
PC1 (Valence): -1.42 Negative
PC2 (Disposition): -0.16

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.330
L0.300
🤔0.296
S0.273
que0.264
Suppressed:
மகி-0.227
rest-0.222
ecstatic-0.212
alleviate-0.210
uplifting-0.196