MiMo-V2.6-Pro
XiaomiThroughout RL, environment hardening, adversarial screening, and verifier cross-checks keep the loop honest against reward hacking.
specific method · filed under post-training
Cross-checks verifiers during RL to help guard against reward hacking.
The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See reward modelling for the methods filed beside it.
Also called verifier cross-checks.
2 quotes from 2 sources, one card per model, each opening with its strongest.
Throughout RL, environment hardening, adversarial screening, and verifier cross-checks keep the loop honest against reward hacking.
Throughout RL, environment hardening, adversarial screening, and verifier cross-checks keep the loop honest against reward hacking.