Training large language models on narrow tasks can lead to broad misalignment
Read Full Article
#AIMisalignment #NarrowTasksBroadImpact #LanguageModelTraining #SafetyFirst #RLHFConcerns
Read Full Article
#AIMisalignment #NarrowTasksBroadImpact #LanguageModelTraining #SafetyFirst #RLHFConcerns
Nature
Training large language models on narrow tasks can lead to broad misalignment
Nature - Finetuning a large language model on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding.
❤1