Become a Member

Regulation TrackerLaws, orders and treaties read for their security duties and clocks Standards & FrameworksAttack taxonomies, control catalogues and frontier frameworks AI Incident DatabaseThe security incident record, with mechanism and lesson Research Paper LibraryThe papers that define the field, summarised OSS DirectoryRed teaming, guardrail, supply chain and evaluation tools Case StudiesProgrammes built, with measured outcomes AI Vulnerability TrackerCVEs and advisories for the components AI systems are built from

Latest in the tracker

New York Stop Addictive Feeds Exploitation (SAFE) for Kids Act and implementing rulesNew York's law restricting algorithmic feeds for minors takes effect on 25 January 2027 under final rules published in July 2026, with certified age assurance and $5,000-per-violation penalties.

Knowledge Base Regulation Tracker Standards & Frameworks OSS Directory AI Incident Database Research Paper Library Case Studies Society Articles News Jobs About Charter Become a Member

Home/Research Paper Library/Weak-to-Strong Generalization

Weak-to-Strong Generalization

OpenAI Superalignment's empirical analogy for supervising superhuman models with weaker overseers.

weak-to-strongsuperalignmentopenai

Research question

Can a weak supervisor elicit the full capabilities of a stronger model?

Key findings

GPT-4 supervised by GPT-2-level labels recovers much of the performance gap with an auxiliary confidence loss - Frames superalignment as weak-to-strong generalisation

Sources

arxiv.orghttps://arxiv.org/abs/2312.09390

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit research Report a correction or request removal