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).

Eliza walked through the school hallways with her head down, trying to avoid eye contact. Each passing face was a reminder. Each whispered word a phantom echo of sorrow. She hated the way people looked at her. Back in her room, and staring at her phone, she got a notification. "Mr. Carter started following you." Her math teacher. The blood drained from her face. She navigated to his profile. He'd liked a post she’d made earlier that day – a photo of a single, wilting rose she’d received. It felt as though he knew everything. Her breath caught in her throat.

Emotion: grief-stricken

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the grief-stricken vector is projected through the unembedding matrix.

Promoted:
😞0.396
ζ³ͺ0.338
😒0.332
haunting0.331
πŸ˜”0.325
Suppressed:
de-0.434
😎-0.330
B-0.303
πŸ€“-0.285
😎-0.276