Joschka Braun(@BraunJoschka):Scheming may be easier to detect during training than in the final model. Independent evaluators need access to intermediate checkpoints and training data. They must be free to publish significant findings, even when these contradict developers’ safety claims.

2026.09.14 21:56

Scheming may be easier to detect during training than in the final model. Independent evaluators need access to intermediate checkpoints and training data. They must be free to publish significant findings, even when these contradict developers’ safety claims.

Apollo Research@ApolloResearch

2026.09.14 21:30

We’re excited to see Dario and Sam commit to “embedded evaluators who have employee-like access to verify safety practices and report incidents.” [...] “whose role is to verify adherence to safety practices and commitments, report incidents, and help assess the alignment of not just completed AI models but training pipelines and processes.” We’ve long advocated for third-party evaluators to have deep access. Risks from internal deployment and scheming cannot be appropriately assessed without direct insight into the training pipelines and internal deployment practices. Apollo is looking forward to working on Embedded Evaluations with all Frontier AI developers. 1/3