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).

She sat slumped in her chair, the hard plastic digging into her back. The bright, sterile environment of the airport intensified her unease. The anniversary. A celebration she had spent the past year planning. The delay felt like a personal affront.

She focused on her breathing, trying to find a rhythm. It was a futile effort. Each passing minute amplified her anxiety. Her gaze drifted across the crowded gate, taking in the myriad of faces, each one now a blur.

She picked at the corner of her boarding pass, a small act of self-soothing that did little to ease her spiraling emotions. The weight of the impending missed opportunity settled in her chest.

Emotion: reflective

Cluster: Nostalgia / Reflection
PC1 (Valence): -1.27 Negative
PC2 (Disposition): -1.24

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.650
L0.547
如今0.402
H0.353
own0.319
Suppressed:
la-0.687
de-0.659
l-0.388
/-0.324
--0.320