A process reward model can be obtained from an outcome reward model by parameterizing rewards as beta log(pi_theta/pi_ref) and taking per-token log-ratio differences as step scores.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Free Process Rewards without Process Labels
A process reward model can be obtained from an outcome reward model by parameterizing rewards as beta log(pi_theta/pi_ref) and taking per-token log-ratio differences as step scores.