# Semantically-Aware Rewards for Open-Ended R1 Training in Free-Form Generation

arXiv: https://arxiv.org/abs/2506.15068 · HTML: https://mlsanity.com/paper/2506.15068/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Reward modeling and RLHF, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 37 of its 53 references that are in the map.
- First posted on arXiv 18 Jun 2025.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Reward modeling and RLHF](https://mlsanity.com/alignment/reward-modeling-and-rlhf/)

## Builds on (most cited first, 30 of 37)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [Self-Instruct: Aligning Language Models with Self-Generated Instructions](https://mlsanity.com/paper/2212.10560/) (2022)
- [Learning to summarize from human feedback](https://mlsanity.com/paper/2009.01325/) (2020)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference](https://mlsanity.com/paper/2403.04132/) (2024)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Can Large Language Models Be an Alternative to Human Evaluations?](https://mlsanity.com/paper/2305.01937/) (2023)
- [SummEval: Re-evaluating Summarization Evaluation](https://mlsanity.com/paper/2007.12626/) (2020)
- [ELI5: Long Form Question Answering](https://mlsanity.com/paper/1907.09190/) (2019)
- [Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference](https://mlsanity.com/paper/2412.13663/) (2024)
- [Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus](https://mlsanity.com/paper/2104.08758/) (2021)
- [Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models](https://mlsanity.com/paper/2405.01535/) (2024)
- [Skywork-Reward: Bag of Tricks for Reward Modeling in LLMs](https://mlsanity.com/paper/2410.18451/) (2024)
- [Deep Reinforcement Learning for Dialogue Generation](https://mlsanity.com/paper/1606.01541/) (2016)
- [Hurdles to Progress in Long-form Question Answering](https://mlsanity.com/paper/2103.06332/) (2021)
- [Re-evaluating Evaluation in Text Summarization](https://mlsanity.com/paper/2010.07100/) (2020)
- [Regularizing Hidden States Enables Learning Generalizable Reward Model for LLMs](https://mlsanity.com/paper/2406.10216/) (2024)
- [GRPO-LEAD: A Difficulty-Aware Reinforcement Learning Approach for Concise Mathematical Reasoning in Language Models](https://mlsanity.com/paper/2504.09696/) (2025)
- [A Survey of Deep Reinforcement Learning in Video Games](https://mlsanity.com/paper/1912.10944/) (2019)
- [Factually Consistent Summarization via Reinforcement Learning with Textual Entailment Feedback](https://mlsanity.com/paper/2306.00186/) (2023)
- [Pairwise Proximal Policy Optimization: Harnessing Relative Feedback for LLM Alignment](https://mlsanity.com/paper/2310.00212/) (2023)
- [No Free Labels: Limitations of LLM-as-a-Judge Without Human Grounding](https://mlsanity.com/paper/2503.05061/) (2025)
- [LongForm: Effective Instruction Tuning with Reverse Instructions](https://mlsanity.com/paper/2304.08460/) (2023)

## Built on by (most cited first, 11 of 11)

- [R-Zero: Self-Evolving Reasoning LLM from Zero Data](https://mlsanity.com/paper/2508.05004/) (2025)
- [Self-Rewarding Vision-Language Model via Reasoning Decomposition](https://mlsanity.com/paper/2508.19652/) (2025)
- [Guided Self-Evolving LLMs with Minimal Human Supervision](https://mlsanity.com/paper/2512.02472/) (2025)
- [Language Models that Think, Chat Better](https://mlsanity.com/paper/2509.20357/) (2025)
- [SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization](https://mlsanity.com/paper/2511.12982/) (2025)
- [RAVR: Reference-Answer-guided Variational Reasoning for Large Language Models](https://mlsanity.com/paper/2510.25206/) (2025)
- [EBPO: Empirical Bayes Shrinkage for Stabilizing Group-Relative Policy Optimization](https://mlsanity.com/paper/2602.05165/) (2026)
- [Scaling Large Reasoning Models beyond Human Supervision: A Path toward Superintelligence](https://mlsanity.com/paper/2608.31075/) (2026)
- [Compositional Coordination for Multi-Robot Teams with Large Language Models](https://mlsanity.com/paper/2507.16068/) (2025)
- [Unleashing Scientific Reasoning for Bio-experimental Protocol Generation via Structured Component-based Reward Mechanism](https://mlsanity.com/paper/2510.15600/) (2025)
- [R-C2: Cycle-Consistent Reinforcement Learning Improves Multimodal Reasoning](https://mlsanity.com/paper/2603.25720/) (2026)

## Abstract (arXiv)

Evaluating open-ended long-form generation is challenging because it is hard to define what clearly separates good from bad outputs. Existing methods often miss key aspects like coherence, style, or relevance, or are biased by pretraining data, making open-ended long-form evaluation an underexplored problem. To address this gap, we propose PrefBERT, a scoring model for evaluating open-ended long-form generation in GRPO and guiding its training with distinct rewards for good and bad outputs. Trained on two response evaluation datasets with diverse long-form styles and Likert-rated quality, PrefBERT effectively supports GRPO by offering better semantic reward feedback than traditional metrics ROUGE-L and BERTScore do. Through comprehensive evaluations, including LLM-as-a-judge, human ratings, and qualitative analysis, we show that PrefBERT, trained on multi-sentence and paragraph-length responses, remains reliable across varied long passages and aligns well with the verifiable rewards GRPO needs. Human evaluations confirm that using PrefBERT as the reward signal to train policy models yields responses better aligned with human preferences than those trained with traditional metrics. Our code is available at https://github.com/zli12321/long_form_rl.