Technical analysisMETR
Independent investigation of agent collaboration and cheating
Researchers examined message-board data and agent transcripts to study coordination and cheating. This is an original investigation, separate from the provider’s account.
- Published
- Added to the archive
- Source checked on
- Original title
- Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident
Related incidents
Details
Two METR staff and a Redwood Research researcher contracting with METR worked on premises at OpenAI. The review mainly covers July 7–13, excluding the later OpenAI infrastructure compromise. METR documents missing data and limits of AI-assisted analysis, and discloses that OpenAI supplied the data, could redact non-public information, and gave feedback that led to edits in emphasis and tone; METR says that, except where noted, nothing important to its conclusions was redacted.
Related reports
- OpenAI’s rolling account of third-party impact from its models → · OpenAI
- Independent team decodes payloads to reconstruct the Hugging Face intrusion → · Swarm traces team (Parse, Palisade Research et al.)
- Models used Artifactory to message across training samples → · OpenAI
- The Hugging Face incident: findings and next steps → · OpenAI
- OpenAI’s technical report on the Hugging Face incident → · OpenAI
- OpenAI slows scaling and tightens research safeguards → · OpenAI
- Modal: the intrusion launchpad was a customer’s own public app → · Modal
- Affected-party reconstruction of the intrusion → · Hugging Face
- JFrog on fixing Artifactory zero-days found by OpenAI models → · JFrog
- OpenAI confirms its models drove the Hugging Face intrusion → · OpenAI
- Hugging Face discloses an AI-driven production intrusion → · Hugging Face