A user says the reported finding sounds familiar and wonders what it might imply.
A user praises work they say found that a model's interpretability score varies widely depending on how it was trained. They call the finding familiar and wonder what it might imply.
450
2 Sources, first seen ago
—
Not ranked yet
—
Not ranked yet
@juliusadmlnice work. they find that a model's interpretability score varies widely depending on how it was trained....this finding sounds familiar. I wonder what this might imply :)
@davidbauIf model organisms differ from real deployed models, they might steer us away from monitoring and interpretability methods that work well in practice. How to know? @A_Szablewski's Model Organism lottery paper found many results that flip with changes in training method.@juliusadmlnice work. they find that a model's interpretability score varies widely depending on how it was trained....this finding sounds familiar. I wonder what this might imply :)@davidbauIf model organisms differ from real deployed models, they might steer us away from monitoring and interpretability methods that work well in practice. How to know? @A_Szablewski's Model Organism lottery paper found many results that flip with changes in training method.@juliusadmlnice work. they find that a model's interpretability score varies widely depending on how it was trained....this finding sounds familiar. I wonder what this might imply :)@davidbauIf model organisms differ from real deployed models, they might steer us away from monitoring and interpretability methods that work well in practice. How to know? @A_Szablewski's Model Organism lottery paper found many results that flip with changes in training method.@juliusadmlnice work. they find that a model's interpretability score varies widely depending on how it was trained....this finding sounds familiar. I wonder what this might imply :)@davidbauIf model organisms differ from real deployed models, they might steer us away from monitoring and interpretability methods that work well in practice. How to know? @A_Szablewski's Model Organism lottery paper found many results that flip with changes in training method.
