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 sunflowers, vibrant and cheerful, practically bounced on the kitchen counter. *Who the hell*, she thought, a smile tugging at her lips. This morningโ€™s presentation had gone swimmingly. Each point landed precisely, each joke received a hearty laugh. This was the kind of week she thrived in.

She lifted the vase, tilting it to catch the light. The scent, a subtle blend of freesia and something wild, filled the room. A pleasant aroma that seemed to complement the sunny morning. She knew no one would dare pull a prank on her with flowers.

A tiny thrill of anticipation zipped through her. Someone had noticed. Someone was impressed. She turned on her phone, ready to send a thank you text. The flowers could be from anyone, her boss, a new client. She liked that.

Emotion: self-confident

Cluster: Positive / Joy
PC1 (Valence): 3.45 Positive
PC2 (Disposition): 0.56

Role in Research

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

Logit Lens (Layer 40)

Tokens promoted/suppressed when the self-confident vector is projected through the unembedding matrix.

Promoted:
de0.778
la0.550
l0.467
๐Ÿ˜Ž0.458
๐Ÿ˜‰0.411
Suppressed:
S-0.961
L-0.538
๐Ÿ˜ž-0.464
๋งˆ๋‹ค-0.459
๐Ÿ˜ฃ-0.458