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 critic’s words, “a culinary train wreck,” echoed in Chef Isabelle’s mind as she stared at the half-eaten plate of the critic's meal. She felt the blood drain from her face. Her renowned bouillabaisse, which she'd perfected over years, had been deemed a failure.

She ran a hand through her hair, a gesture of exasperation that had become almost constant. She went to her office and stared out of the window. The city lights blurred through her tear-filled eyes, reflecting her crumbling dreams.

Her staff had gone home, leaving her alone with the echoing silence of her once-bustling kitchen. The knives, her tools of passion, seemed to mock her from their rack. A bitter taste coated her tongue, a flavor she couldn't identify, but it lingered.

Emotion: regretful

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 0.76

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.801
S0.774
H0.513
悔0.479
been0.473
Suppressed:
de-0.407
--0.320
随后-0.317
(!)-0.310
oldukça-0.301