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).

He stared, mouth agape, at the brightly colored flyer clutched in his hand. “Karaoke Night – Alex’s 25th Birthday!” it proclaimed, with a photo of Alex, beaming. Mark’s own birthday party, a low-key board game night, was slated for that very evening. A flush crept up his neck.

He sank onto his sofa. He’d told Alex about his plans weeks ago. Alex had seemed… fine with it. They’d even talked about maybe grabbing a celebratory pizza together *afterwards*. The pizza felt a long way off.

He felt a prickling sensation behind his eyes. He’d spent so long prepping for this, buying new board games, curating a playlist. He’d envisioned a relaxed, fun evening. Now, the evening felt anything but.

Emotion: surprised

Cluster: Surprise / Confusion
PC1 (Valence): -0.58 Negative
PC2 (Disposition): -0.18

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
que0.495
de0.447
la0.431
l0.323
😳0.296
Suppressed:
每天-0.269
cheerfully-0.249
ing-0.238
alleviate-0.224
gladly-0.219