arXiv preprint ·
Many Right Answers, One Training Example: Removing Responder Noise from GRPO for LLM Recommenders
- Published
- Source
- arXiv preprint, posted before peer review
- Original
- Read it at arXiv
- Topics
Abstract
A recommendation request on Reddit draws several replies, each naming different titles. When an LLM recommender is trained on such threads with reinforcement learning (RL), which replies should a generated list be rewarded against? The state-of-the-art recipe, Rank-GRPO, picks one reply per training example, so the advantage mostly records which reply was drawn (85% of its variance) rather than how good the list is. We argue that RL should train on the request, not on each answer, and propose ReMaRk: it scores every list once against the union of all replies' titles and samples each request equally. This makes RL cheaper. On seven backbones and three datasets, ReMaRk reaches the validation recall of 2,000 per-reply RL steps in 25 to 53% of the steps and leads at that budget; trained until neither improves, it is never significantly behind Rank-GRPO and in four of seven pairs reaches Rank-GRPO's best with fewer GPU-hours. The same reward helps on ProtoQA, where about 100 people answer each question, which points to a simple rule for RL with several valid human answers: reward them jointly.
The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.
Read it on ReadSearch
Each opens after you create an account, and brings you back to this paper.
More in Computation & Language / NLP
Everyone Was Right: Lexical Representations in Language Models Are Both Abstract and Holistic
TopicMerge: A Divide and Conquer Approach to Topic Modeling
Many Steps into One: Reasoning Contraction and the Limits of Safe Memory Reuse for LLM Agents
LLM-jp-4 Technical Report
Reference Wording Shapes Retrieval of Robot Task Instructions
Get the few papers worth your time in Computation & Language / NLP every morning.
ReadSearch reads the new papers in your field each day and brings you the few worth reading, each with its key finding in a sentence.