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 late afternoon sun cast long shadows across Amelia's overgrown lawn. She stared out the kitchen window, the warmth doing nothing to penetrate the chill that had settled deep within her bones. A knock at the back door startled her. It was Mr. Henderson, his usual jovial face a little hesitant. He wanted to talk about something involving her unused patch of land beside his property.

"I was wondering," he began, fidgeting with the brim of his hat, "if I might... well, perhaps borrow a bit of your yard. To, you know, grow some vegetables." Amelia just nodded, her throat tight.

The thought of someone else tending to the space, bringing life and activity where she felt only emptiness, felt like a small, sharp pinprick. She mumbled an assent, a choked sound lost in the rustling of leaves. Mr. Henderson beamed, thanked her profusely, and turned back towards his own home, leaving Amelia alone with the growing dusk and a heavy sensation in her chest.

Emotion: sad

Cluster: Sadness / Despair
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.436
😔0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322