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 house seemed to shrink with each passing day. The air hung thick with memories, each one a sharp stab to the heart. Carol felt the familiar weight of despair, the heavy cloak she couldnโ€™t seem to shrug off. She was looking through Markโ€™s belongings, something she had been putting off for weeks. She knew she had to. There it was: a small, unassuming envelope slipped inside the back cover of his favorite book, a bank statement.

Her breath hitched as she read the figures. A sudden wave of nausea swept over her, causing her to clutch her stomach. She felt a sharp, unexpected pain, a physical manifestation of her sorrow. She sat down abruptly, the book falling from her grasp, the pages splayed open. She stared blankly at the ceiling, feeling an overwhelming sense of isolation.

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:
๐Ÿ˜ž0.396
ๆณช0.338
๐Ÿ˜ข0.332
haunting0.331
๐Ÿ˜”0.325
Suppressed:
de-0.434
๐Ÿ˜Ž-0.330
B-0.303
๐Ÿค“-0.285
๐Ÿ˜Ž-0.276