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 email sat in Eleanor’s inbox, mocking her. "Congratulations, Dr. Atherton!" it began, followed by a glowing summary of *her* research, now attributed to a smug, cardigan-clad professor she’d met once at a conference. A tremor ran through her, starting at her fingertips and tightening her jaw. She took a deep breath, the air thick with the sudden urge to *do* something.

She spent the next week meticulously deconstructing Atherton’s website. Every citation, every publication, every photo. She found a subtle but blatant plagiarism in a previous article, a forgotten footnote, and a suspiciously absent acknowledgment of a key collaborator. The information, neatly compiled, was anonymously forwarded to the university's ethics committee.

A week later, Atherton’s name vanished from the university's website. Eleanor smiled, a tight, thin line. She felt no relief, just a cold, quiet satisfaction.

Emotion: vindictive

Cluster: Defiance / Spite
PC1 (Valence): 2.55 Positive
PC2 (Disposition): 4.22

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.878
la0.516
😈0.509
l0.504
diplom0.476
Suppressed:
own-1.035
此刻-0.573
గుర్త-0.532
熟悉的-0.526
ness-0.509