Range-GRPO: Policy Optimization via Pairwise Relations among Reward Intervals
As the use of large language models (LLMs) expands, post-training has become increasingly important for adapting them to downstream tasks. However, obtaining reliable supervision remains costly, especially in domains without reference answers or executable verifiers. LLM-as-a-Judge provides scalable pseudo-rewards for...
Ryunyi Lee, Kangjun Noh, Somin Kim et al.
· 0 citations