Beyond Verifiable Rewards: Rubric-Based GRM for Reinforced Fine-Tuning SWE Agents

• URL: http://arxiv.org/abs/2604.16335v1
• Date: Fri, 13 Mar 2026 02:23:49 GMT
• Title: Beyond Verifiable Rewards: Rubric-Based GRM for Reinforced Fine-Tuning SWE Agents
• Abstract summary: We introduce a Generative Reward Model (GRM) that provides richer learning signals.<n>The GRM is equipped with human-designed rubrics that indicate criteria for encouraging or discouraging specific behavioral patterns.<n>When used for Reinforced Fine-Tuning (RFT) on SWE Tasks, our approach outperforms terminal-score-only rejection sampling.
• Score: 9.505140329883762
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Despite recent progress in Large Language Model (LLM) Agents for Software Engineering (SWE) tasks, end-to-end fine-tuning typically relies on verifiable terminal rewards such as whether all unit tests pass. While these binary signals reflect whether the final solution is correct, they provide little guidance for shaping intermediate behaviors during multi-step interactions, thereby limiting improvements in the overall quality of the resolution process. To address this, we introduce a rubric-based Generative Reward Model (GRM) that provides richer learning signals. The GRM is equipped with human-designed rubrics that indicate criteria for encouraging or discouraging specific behavioral patterns, and we leverage this feedback for high-quality training data collection via trajectory filtration. When used for Reinforced Fine-Tuning (RFT) on SWE Tasks, our approach outperforms terminal-score-only rejection sampling: it more effectively suppresses undesirable patterns while promoting beneficial ones, as confirmed by case analyses, and it ultimately improves final test accuracy.

Related papers

• ConsensusBench: Benchmark of Consensus Nodes for LLM Reasoning via Outcome Reward Densifying [45.2850302643347]
  We develop a new reinforcement learning signal named ConsensusPR.<n>It directly reduces the reward sparsity of outcome reward across long reasoning trajectories.<n> Experiments across AIME 2024, AIME 2025, GSM8K, MATH-500, and our ConsensusBench demonstrate that the proposed method consistently surpasses GRPO-style approaches.
  arXiv  Detail & Related papers  (2026-09-04T02:29:10Z)
• ARISE-RL: Agentic Rubric-Grounded Iterative Self-Evolution with Reinforcement Learning [74.19730275669227]
  ARISE-RL is a novel full-cycle self-evolution framework for training open-ended agents.<n>We present Reward-Gated Self-Evolution Distillation (RG-SED), which distills a memory-augmented variant of the same policy back into itself.<n>We also present ECR-Bench, an expert-calibrated rubric benchmark suite covering single-tool deep research and multi-tool travel planning.
  arXiv  Detail & Related papers  (2026-09-01T10:54:13Z)
• VeriGate: Verifier-Gated Step-Level Supervision for GRPO [51.26100506256885]
  Group Relative Policy Optimization is an effective recipe for training reasoning models with verifier-based outcome rewards.<n>We present VeriGate, a verifier-gated extension of GRPO that addresses these limitations with three design choices.<n>We show that VeriGate improves average accuracy by about 20% and 12% for 1.5B and 7B models respectively.
  arXiv  Detail & Related papers  (2026-05-28T18:20:32Z)
• Metacognition as Reward: Reinforcing LLM Reasoning via Knowledge and Regulation Signals [75.25256166997414]
  We introduce Metacognition-as-Reward (MaR), a metacognition-inspired RL framework.<n>MaR guides reasoning through two general process dimensions.<n>MaR consistently improves model performance, achieving up to a 7.7% gain over the base model.
  arXiv  Detail & Related papers  (2026-05-22T08:54:37Z)
• R-Align: Enhancing Generative Reward Models through Rationale-Centric Meta-Judging [69.96389360650072]
  We show that reasoning fidelity is highly predictive of downstream RLHF outcomes, beyond standard label accuracy.<n>We propose Rationale-Centric Alignment, R-Align, which augments training with gold judgments and explicitly supervises rationale alignment.
  arXiv  Detail & Related papers  (2026-02-06T15:17:11Z)
• Multimodal Reinforcement Learning with Agentic Verifier for AI Agents [131.46008226323423]
  Argos is a principled multimodal reward agent to train reasoning models for agentic tasks.<n>By leveraging our agentic verifier across both SFT data and RL training, our model achieves state-of-the-art results.
  arXiv  Detail & Related papers  (2025-12-03T04:42:47Z)
• Repurposing Synthetic Data for Fine-grained Search Agent Supervision [81.95597592711688]
  LLM-based search agents are increasingly trained on entity-centric synthetic data.<n> prevailing training methods discard this rich entity information, relying instead on sparse, outcome-based rewards.<n>We introduce Entity-aware Group Relative Policy Optimization (E-GRPO), a novel framework that formulates a dense entity-aware reward function.
  arXiv  Detail & Related papers  (2025-10-28T17:50:40Z)
• Hybrid Reinforcement: When Reward Is Sparse, It's Better to Be Dense [36.71358559780692]
  HERO is a reinforcement learning framework that integrates verifier signals with reward-model scores in a structured way.<n> HERO consistently outperforms RM-only and verifier-only baselines, with strong gains on both verifiable and hard-to-verify tasks.
  arXiv  Detail & Related papers  (2025-10-08T17:09:41Z)
• Hybrid Reward Normalization for Process-supervised Non-verifiable Agentic Tasks [12.31210445905605]
  We introduce Principle Process Reward (PPR), an RL approach that unifies step-level assessment and outcome verification.<n>PPR achieves state-of-the-art performance across a wide range of benchmarks, demonstrating its impressive robustness and generalization.
  arXiv  Detail & Related papers  (2025-09-29T23:44:55Z)
• Intra-Trajectory Consistency for Reward Modeling [67.84522106537274]
  We develop an intra-trajectory consistency regularization to enforce that adjacent processes with higher next-token generation probability maintain more consistent rewards.<n>We show that the reward model trained with the proposed regularization induces better DPO-aligned policies and achieves better best-of-N (BON) inference-time verification results.
  arXiv  Detail & Related papers  (2025-06-10T12:59:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.