LLM safety training fails due to competing objectives and mismatched generalization, enabling new jailbreaks that succeed on all unsafe prompts from red-teaming sets in GPT-4 and Claude.
Pretraining language models with human preferences
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 4roles
background 1polarities
background 1representative citing papers
Self-Debugging teaches LLMs to identify and fix their own code errors through rubber-duck-style natural language explanations and execution feedback, delivering 2-12% gains over baselines on Spider, TransCoder, and MBPP.
Self-generated replay from language models nearly eliminates catastrophic forgetting during finetuning except when models are pretrained close to saturation.
PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.
citing papers explorer
-
Jailbroken: How Does LLM Safety Training Fail?
LLM safety training fails due to competing objectives and mismatched generalization, enabling new jailbreaks that succeed on all unsafe prompts from red-teaming sets in GPT-4 and Claude.