Chaining the Evidence: Robust Reinforcement Learning for Deep Search Agents with Citation-Aware Rubric Rewards

TL;DR

This paper introduces Citation-aware Rubric Rewards (CaRR) and C-GRPO, a reinforcement learning framework that improves deep search agents by emphasizing factual grounding, reasoning comprehensiveness, and evidence connectivity, outperforming standard methods.

Contribution

The paper proposes CaRR and C-GRPO, novel reward and training methods that enhance the factuality and reasoning quality of deep search agents in RL settings.

Findings

C-GRPO outperforms baseline RL methods on multiple benchmarks.

CaRR reduces shortcut exploitation and hallucinations.

Agents trained with C-GRPO generalize well to open-ended tasks.

Abstract

Reinforcement learning (RL) has emerged as a critical technique for enhancing LLM-based deep search agents. However, existing approaches primarily rely on binary outcome rewards, which fail to capture the comprehensiveness and factuality of agents' reasoning process, and often lead to undesirable behaviors such as shortcut exploitation and hallucinations. To address these limitations, we propose \textbf{Citation-aware Rubric Rewards (CaRR)}, a fine-grained reward framework for deep search agents that emphasizes reasoning comprehensiveness, factual grounding, and evidence connectivity. CaRR decomposes complex questions into verifiable single-hop rubrics and requires agents to satisfy these rubrics by explicitly identifying hidden entities, supporting them with correct citations, and constructing complete evidence chains that link to the predicted answer. We further introduce \textbf{Citation-aware Group Relative Policy Optimization (C-GRPO)}, which combines CaRR and outcome rewards for training robust deep search agents. Experiments show that C-GRPO consistently outperforms standard outcome-based RL baselines across multiple deep search benchmarks. Our analysis also validates that C-GRPO effectively discourages shortcut exploitation, promotes comprehensive, evidence-grounded reasoning, and exhibits strong generalization to open-ended deep research tasks. Our code and data are available at https://github.com/THUDM/CaRR.