From the article, seems like TOPLOC:
> based on top of novel components such as TOPLOC, which verifies rollouts from untrusted inference workers
> based on top of novel components such as TOPLOC, which verifies rollouts from untrusted inference workers
At a glance it looks like something akin to a computing a checksum that's locality sensitive, so it's robust to floating point errors, etc.
What's to stop someone from sending bad data + a matching bad checksum?
The checksum is validated by redoing the computation, but making use of the fact that you already have the entire response to enable greater parallelism than when generating it one token at a time.