logseq/pages/hls__Data_Poisoning_in_LLMs_1709648003419_0.md

file:: Data_Poisoning_in_LLMs_1709648003419_0.pdf file-path:: ../assets/Data_Poisoning_in_LLMs_1709648003419_0.pdf

• in particular, ls-type:: annotation hl-page:: 1 hl-color:: red id:: 65e728b0-0b5a-4a1a-8f0f-5d78e2e460fc
• It is not possible to adequately curate the training data to assure data quality at this massive scale. ls-type:: annotation hl-page:: 1 hl-color:: green id:: 65e729aa-c490-4bbb-b797-52837921787f
• simple to contaminate small bits of data ls-type:: annotation hl-page:: 1 hl-color:: yellow id:: 65e729d8-b092-4534-8ec4-14b02e5f7856