{"id":"arxiv:2311.14455","kind":"preprint","title":"Universal Jailbreak Backdoors from Poisoned Human Feedback","authors":["Javier Rando","Florian Tramèr"],"author_count":2,"year":2023,"published":"2023-11-24","venue":null,"cited_by":null,"doi":null,"arxiv_id":"2311.14455","nct_id":null,"source_url":"https://arxiv.org/abs/2311.14455","archive_url":"https://space-kitty.com/papers/arxiv:2311.14455","updated":"2024-04-29","pmid":null,"pmcid":null,"openalex_id":null,"pdf_url":"https://arxiv.org/pdf/2311.14455","abstract":null,"abstract_license":null,"license":null,"is_open_access":false,"retracted":false,"peer_reviewed":false,"categories":["cs.AI","cs.CL","cs.CR","cs.LG"],"details":null,"provenance":"arxiv","attribution":null,"topics":[],"cited_in_dossiers":[]}