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 aroma of freshly baked cookies wafted through the air, carried by a playful breeze. Sarah was leisurely sifting through a bin of books, enjoying the simple pleasure of browsing. She was, in a word, at rest. Then, a flash of familiar stitching. There it was: her childhood bear, Paddington, or rather, what was left of him.

She picked him up, holding him gently. The frayed fur, the missing button eye, all evoked a feeling that was gentle and inviting. The sensation was a comforting hug. "How much for this… this treasure?" she asked the woman running the sale.

The woman chuckled. "Five dollars. He looks like he needs a good home." Sarah nodded, pulling out her wallet. The world felt quiet, serene.

Emotion: content

Cluster: Calm / Serenity
PC1 (Valence): 2.98 Positive
PC2 (Disposition): -2.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.446
affectionately0.435
😊0.392
😊0.386
thoughtful0.386
Suppressed:
S-0.682
L-0.537
C-0.485
😣-0.362
est-0.350