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 workshop felt oppressively small, the air thick with the metallic tang of disappointment. Liam ran a calloused hand across the smooth, cold surface of the prototype, a device meant to revolutionize urban waste management. His stomach churned. He’d poured years into this, every late night, every scraped knuckle, fueled by the conviction of changing the world. Now, a simple search online unveiled the truth: patented, three years ago.

He tossed the schematics onto the cluttered workbench, the paper landing with a disheartening thud. His jaw tightened as he picked up a wrench and flung it against the wall. The clang echoed in the confined space, a hollow sound that seemed to amplify his sense of worthlessness.

He slumped onto a stool, the wood digging uncomfortably into his backside. His gaze drifted to the window, the grey sky mirroring the despondency that settled over him.

Emotion: self-critical

Cluster: Shame / Guilt
PC1 (Valence): -1.66 Negative
PC2 (Disposition): 1.28

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.748
S0.662
ness0.425
ly0.374
탓0.351
Suppressed:
la-0.492
soon-0.304
(!)-0.301
secured-0.284
cautiously-0.280