Comment by Anna Wang

“Before a new model can be deployed internally, you must first test it for a minimum, fixed period. [...] A model family becomes covered as soon as any of its checkpoints (pre-, mid- or post-trained) is substantially more capable than the last model to complete its wait. [...] There is also an absolute floor on capabilities, so that this doesn’t apply to lots of benign models. [...] The developer continually pulls checkpoints throughout the training process. They first send an auditor a hash, a short fingerprint of the exact weights; then, they carry out safety testing for X days before deploying a checkpoint broadly internally without the below rules. All work has to be fully transparent to auditors. [...] Stage 2, second part of X: broad monitored internal use, but the model can’t launch jobs, merge code, write to training data, or touch prod.”