RL for LLM Agents

RL: Reinforcement Learning · LLM: Large Language Model

Ask about this topic

Overview

This field studies how to train large language model agents to act over multiple steps in environments such as web browsers, coding tools, games, and simulated workplaces. Common approaches use policy-gradient methods such as PPO and GRPO, verifiable task rewards, and process-level credit assignment to connect outcomes to particular actions or turns. Other methods teach agents to plan hierarchically, manage memory and reusable skills, or coordinate multiple agents, often using simulators and environment-generated feedback. A central challenge is making learning stable and efficient when rewards are sparse and tasks are long, while ensuring gains transfer to new environments and do not come from exploiting flawed reward signals.

Latest papers

CardsList

616 papers134 papers in the last four weeks, up 139% on the four weeks before. 0.8% of all new papers.

Apr 2, 2026

ProCeedRL: Process Critic with Exploratory Demonstration Reinforcement Learning for LLM Agentic Reasoning

Reinforcement Learning (RL) significantly enhances the reasoning abilities of large language models (LLMs), yet applying it to multi-turn agentic tasks remains challenging due to the long-horizon nature of interactions and the stochasticity of environmental feedback. We identify a structural failure mode in agentic exploration: suboptimal actions elicit noisy observations into misleading contexts, which further weaken subsequent decision-making, making recovery increasingly difficult. This cumulative feedback loop of errors renders standard exploration strategies ineffective and susceptible to the model's reasoning and the environment's randomness. To mitigate this issue, we propose ProCeedRL: Process Critic with Explorative Demonstration RL, shifting exploration from passive selection to active intervention. ProCeedRL employs a process-level critic to monitor interactions in real time, incorporating reflection-based demonstrations to guide agents in stopping the accumulation of errors. We find that this approach significantly exceeds the model's saturated exploration performance, demonstrating substantial exploratory benefits. By learning from exploratory demonstrations and on-policy samples, ProCeedRL significantly improves exploration efficiency and achieves superior performance on complex deep search and embodied tasks.

Jingyue Gao, Yanjiang Guo, Xiaoshuai Chen +1

Institute for Interdisciplinary Information Sciences, Tsinghua University · Shanghai Qizhi Institute · Independent Researcher

Agentic RLRL for Language Model ReasoningRL for LLM AgentsRL ExplorationProcess Reward Models
AskPDF
Apr 2, 2026

ContextBudget: Budget-Aware Context Management for Long-Horizon Search Agents

LLM-based agents show strong potential for long-horizon reasoning, yet their context size is limited by deployment factors (e.g., memory, latency, and cost), yielding a constrained context budget. As interaction histories grow, this induces a trade-off between retaining past information and staying within the context limit. To address this challenge, we propose Budget-Aware Context Management (BACM), which formulates context management as a sequential decision problem with a context budget constraint. It enables agents to assess the available budget before incorporating new observations and decide when and how much of the interaction history to compress. We further develop BACM-RL, an end-to-end curriculum-based reinforcement learning approach that learns compression strategies under varying context budgets. Experiments on compositional multi-objective QA and long-horizon web browsing benchmarks show that BACM-RL consistently outperforms prior methods across model scales and task complexities, achieving over 1.6×1.6\times gains over strong baselines in high-complexity settings, while maintaining strong advantages as budgets shrink, where most methods exhibit a downward performance trend.

Yong Wu, YanZhao Zheng, TianZe Xu +9

Zhejiang University, Hangzhou, China · Alibaba Group, Hangzhou, China

LLM Context ManagementLong-Horizon LLM AgentsContext CompressionRL for LLM Agents
AskPDF
Apr 2, 2026

DeltaMem: Towards Agentic Memory Management via Reinforcement Learning

Recent advances in persona-centric memory have revealed the powerful capability of multi-agent systems in managing persona memory, especially in conversational scenarios. However, these complex frameworks often suffer from information loss and are fragile across varying scenarios, resulting in suboptimal performance. In this paper, we propose DeltaMem, an agentic memory management system that formulates persona-centric memory management as an end-to-end task within a single-agent setting. To further improve the performance of our agentic memory manager, we draw inspiration from the evolution of human memory and synthesize a user-assistant dialogue dataset along with corresponding operation-level memory updating labels. Building on this, we introduce a novel Memory-based Levenshtein Distance to formalize the memory updating reward, and propose a tailored reinforcement learning framework to further enhance the management capabilities of DeltaMem. Extensive experiments show that both training-free and RL-trained DeltaMem outperform all product-level baselines across diverse long-term memory benchmarks, including LoCoMo, HaluMem, and PersonaMem.

Qi Zhang, Shen Huang, Chu Liu +4

Alibaba Tongyi Lab, Hangzhou, China · Zhejiang University, Hangzhou, China

Agent Memory ManagementAgent MemoryLong-Term Conversational MemoryRL for LLM AgentsReinforcement Learning
AskPDF
Apr 1, 2026

LangMARL: Natural Language Multi-Agent Reinforcement Learning

Large language model (LLM) agents struggle to autonomously evolve coordination strategies in dynamic environments, largely because coarse global outcomes obscure the causal signals needed for local policy refinement. We identify this bottleneck as a multi-agent credit assignment problem, which has long been studied in classical multi-agent reinforcement learning (MARL) but remains underaddressed in LLM-based systems. Building on this observation, we propose LangMARL, a framework that brings credit assignment and policy gradient evolution from cooperative MARL into the language space. LangMARL introduces agent-level language credit assignment, pioneers gradient evolution in language space for policy improvement, and summarizes task-relevant causal relations from replayed trajectories to provide dense feedback and improve convergence under sparse rewards. Extensive experiments across diverse cooperative multi-agent tasks demonstrate improved sample efficiency, interpretability, and strong generalization.

Huaiyuan Yao, Longchao Da, Xiaoou Liu +3

Arizona State University · Cisco Research · University of North Carolina at Chapel Hill

RL for LLM AgentsMulti-Agent Reinforcement LearningCooperative MARL
AskPDF
Apr 1, 2026

Agent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning

Large Language Models (LLMs) have shown remarkable performance in completing various tasks. However, solving complex problems often requires the coordination of multiple agents, raising a fundamental question: how to effectively select and interconnect these agents. In this paper, we propose Agent Q-Mix, a reinforcement learning framework that reformulates topology selection as a cooperative Multi-Agent Reinforcement Learning (MARL) problem. Our method learns decentralized communication decisions using QMIX value factorization, where each agent selects from a set of communication actions that jointly induce a round-wise communication graph. At its core, Agent Q-Mix combines a topology-aware GNN encoder, GRU memory, and per-agent Q-heads under a Centralized Training with Decentralized Execution (CTDE) paradigm. The framework optimizes a reward function that balances task accuracy with token cost. Across seven core benchmarks in coding, reasoning, and mathematics, Agent Q-Mix achieves the highest average accuracy compared to existing methods while demonstrating superior token efficiency and robustness against agent failure. Notably, on the challenging Humanity's Last Exam (HLE) using Gemini-3.1-Flash-Lite as a backbone, Agent Q-Mix achieves 20.8% accuracy, outperforming Microsoft Agent Framework (19.2%) and LangGraph (19.2%), followed by AutoGen and Lobster by OpenClaw. These results underscore the effectiveness of learned, decentralized topology optimization in pushing the boundaries of multi-agent reasoning.

Eric Hanchen Jiang, Levina Li, Rui Sun +9

University of California Los Angeles · University of Washington · Northwestern University

Multi-Agent Communication TopologyMulti-Agent ReasoningMulti-Agent CommunicationRL for LLM AgentsCooperative MARL
AskPDF
Mar 31, 2026

MemFactory: Unified Inference & Training Framework for Agent Memory

Memory-augmented Large Language Models (LLMs) are essential for developing capable, long-term AI agents. Recently, applying Reinforcement Learning (RL) to optimize memory operations, such as extraction, updating, and retrieval, has emerged as a highly promising research direction. However, existing implementations remain highly fragmented and task-specific, lacking a unified infrastructure to streamline the integration, training, and evaluation of these complex pipelines. To address this gap, we present MemFactory, the first unified, highly modular training and inference framework specifically designed for memory-augmented agents. Inspired by the success of unified fine-tuning frameworks like LLaMA-Factory, MemFactory abstracts the memory lifecycle into atomic, plug-and-play components, enabling researchers to seamlessly construct custom memory agents via a "Lego-like" architecture. Furthermore, the framework natively integrates Group Relative Policy Optimization (GRPO) to fine-tune internal memory management policies driven by multi-dimensional environmental rewards. MemFactory provides out-of-the-box support for recent cutting-edge paradigms, including Memory-R1, RMM, and MemAgent. We empirically validate MemFactory on the open-source MemAgent architecture using its publicly available training and evaluation data. Across the evaluation sets, MemFactory improves performance over the corresponding base models on average, with relative gains of up to 14.8%. By providing a standardized, extensible, and easy-to-use infrastructure, MemFactory significantly lowers the barrier to entry, paving the way for future innovations in memory-driven AI agents.

Ziliang Guo, Ziheng Li, Bo Tang +2

MemTensor

LLM Agent MemoryAgent Memory ManagementLLM Agent TrainingRL for LLM Agents
AskPDF
Mar 25, 2026

MARCH: Multi-Agent Reinforced Self-Check for LLM Hallucination

Hallucination remains a critical bottleneck for large language models (LLMs), undermining their reliability in real-world applications, especially in Retrieval-Augmented Generation (RAG) systems. While existing hallucination detection methods employ LLM-as-a-judge to verify LLM outputs against retrieved evidence, they suffer from inherent confirmation bias, where the verifier inadvertently reproduces the errors of the original generation. To address this, we introduce Multi-Agent Reinforced Self-Check for Hallucination (MARCH), a framework that enforces rigorous factual alignment by leveraging deliberate information asymmetry. MARCH orchestrates a collaborative pipeline of three specialized agents: a Solver, a Proposer, and a Checker. The Solver generates an initial RAG response, which the Proposer decomposes into claim-level verifiable atomic propositions. Crucially, the Checker validates these propositions against retrieved evidence in isolation, deprived of the Solver's original output. This well-crafted information asymmetry scheme breaks the cycle of self-confirmation bias. By training this pipeline with multi-agent reinforcement learning (MARL), we enable the agents to co-evolve and optimize factual adherence. Extensive experiments across hallucination benchmarks demonstrate that MARCH substantially reduces hallucination rates. Notably, an 8B-parameter LLM equipped with MARCH achieves performance competitive with powerful closed-source models. MARCH paves a scalable path for factual self-improvement of LLMs through co-evolution. The code is at https://github.com/Qwen-Applications/MARCH.

Zhuo Li, Yupeng Zhang, Pengyu Cheng +8

Qwen Large Model Application Team, Alibaba · The Chinese University of Hong Kong, Shenzhen

LLM Hallucination MitigationLLM Hallucination DetectionRAG RobustnessBias in LLM-as-a-JudgeMulti-Agent LLMsRL for LLM Agents
AskPDF
Mar 25, 2026

From Pixels to Digital Agents: An Empirical Study on the Taxonomy and Technological Trends of Reinforcement Learning Environments

The remarkable progress of reinforcement learning (RL) is intrinsically tied to the environments used to train and evaluate artificial agents. Moving beyond traditional qualitative reviews, this work presents a large-scale, data-driven empirical investigation into the evolution of RL environments. By programmatically processing a massive corpus of academic literature and rigorously distilling over 2,000 core publications, we propose a quantitative methodology to map the transition from isolated physical simulations to generalist, language-driven foundation agents. Implementing a novel, multi-dimensional taxonomy, we systematically analyze benchmarks against diverse application domains and requisite cognitive capabilities. Our automated semantic and statistical analysis reveals a profound, data-verified paradigm shift: the bifurcation of the field into a "Semantic Prior" ecosystem dominated by Large Language Models (LLMs) and a "Domain-Specific Generalization" ecosystem. Furthermore, we characterize the "cognitive fingerprints" of these distinct domains to uncover the underlying mechanisms of cross-task synergy, multi-domain interference, and zero-shot generalization. Ultimately, this study offers a rigorous, quantitative roadmap for designing the next generation of Embodied Semantic Simulators, bridging the gap between continuous physical control and high-level logical reasoning.

Lijing Luo, Yiben Luo, Alexey Gorbatovski +2

Sun Yat-sen University Shenzhen, China · Yancheng Institute of Technology Yancheng, China · Central University Moscow, Russia +1

RL BenchmarksReinforcement LearningGeneralization in RLRL for LLM AgentsMulti-Task RL
AskPDF
Mar 24, 2026

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

Video understanding with multimodal large language models (MLLMs) remains challenging due to the long token sequences of videos, which contain extensive temporal dependencies and redundant frames. Existing approaches typically treat MLLMs as passive recognizers, processing entire videos or uniformly sampled frames without adaptive reasoning. Recent agent-based methods introduce external tools, yet still depend on manually designed workflows and perception-first strategies, resulting in inefficiency on long videos. We present EVA, an Efficient Reinforcement Learning framework for End-to-End Video Agent, which enables planning-before-perception through iterative summary-plan-action-reflection reasoning. EVA autonomously decides what to watch, when to watch, and how to watch, achieving query-driven and efficient video understanding. To train such agents, we design a simple yet effective three-stage learning pipeline - comprising supervised fine-tuning (SFT), Kahneman-Tversky Optimization (KTO), and Group Relative Policy Optimization (GRPO) - that bridges supervised imitation and reinforcement learning. We further construct high-quality datasets for each stage, supporting stable and reproducible training. We evaluate EVA on six video understanding benchmarks, demonstrating its comprehensive capabilities. Compared with existing baselines, EVA achieves a substantial improvement of 6-12% over general MLLM baselines and a further 1-3% gain over prior adaptive agent methods.

Yaolun Zhang, Ruohui Wang, Jiahao Wang +6

SenseTime Research

Video UnderstandingTemporal Video UnderstandingRL for LLM AgentsReinforcement Learning
AskPDF
Mar 23, 2026

A Context Engineering Framework for Improving Enterprise AI Agents based on Digital-Twin MDP

Despite rapid progress in AI agents for enterprise automation and decision-making, their real-world deployment and further performance gains remain constrained by limited data quality and quantity, complex real-world reasoning demands, difficulties with self-play, and the lack of reliable feedback signals. To address these challenges, we propose a lightweight, model-agnostic framework for improving LLM-based enterprise agents via offline reinforcement learning (RL). The proposed Context Engineering via DT-MDP (DT-MDP-CE) framework comprises three key components: (1) A Digital-Twin Markov Decision Process (DT-MDP), which abstracts the agent's reasoning behavior as a finite MDP; (2) A robust contrastive inverse RL, which, armed with the DT-MDP, to efficiently estimate a well-founded reward function and induces policies from mixed-quality offline trajectories; and (3) RL-guided context engineering, which uses the policy obtained from the integrated process of (1) and (2), to improve the agent's decision-making behavior. As a case study, we apply the framework to a representative task in the enterprise-oriented domain of IT automation. Extensive experimental results demonstrate consistent and significant improvements over baseline agents across a wide range of evaluation settings, suggesting that the framework can generalize to other agents sharing similar characteristics in enterprise environments.

Xi Yang, Aurelie Lozano, Naoki Abe +6

IBM Software Innovation Lab, Yorktown Heights, New York, USA

LLM AgentsLLM Agent TrainingRL for LLM AgentsReinforcement LearningOffline RL
AskPDF
Mar 23, 2026

SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation

Despite the remarkable success of large language models (LLMs), they still face bottlenecks while deploying in dynamic, real-world settings with primary challenges being concept drift and the high cost of gradient-based adaptation. Traditional fine-tuning (FT) struggles to adapt to non-stationary data streams without resulting in catastrophic for getting or requiring extensive manual data curation. To address these limitations within the streaming and continual learning paradigm, we propose the Self-Optimizing Lifelong Autonomous Reasoner (SOLAR) which is an open-ended autonomous agent that leverages parameter-level meta-learning to self-improve, treating model weights as an environment for exploration. It initiates the process by consolidating a strong prior over common-sense knowledge making it effective for transfer-learning. By utilizing a multi-level reinforcement learning approach, SOLAR autonomously discovers adaptation strategies, enabling efficient test-time adaptation to unseen domains. Crucially, SOLAR maintains an evolving knowledge base of valid modification strategies, implicitly acting as an episodic memory buffer to balance plasticity (adaptation to new tasks) and stability (retention of meta-knowledge). Experiments demonstrate that SOLAR outperforms strong baselines on common-sense, mathematical, medical, coding, social and logical reasoning tasks, marking a significant step toward autonomous agents capable of lifelong adaptation in evolving environments.

Nitin Vetcha, Dianbo Liu

Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore · Department of Computational and Data Sciences, Indian Institute of Science, Bangalore, Karnataka, India

Continual Learning for LLM AgentsContinual Learning for LLMsRL for LLM AgentsMeta-LearningTest-Time AdaptationMemory-Augmented Language Model Agents
AskPDF
Mar 22, 2026

PivotRL: High Accuracy Agentic Post-Training at Low Compute Cost

Post-training for long-horizon agentic tasks has a tension between compute efficiency and generalization. While supervised fine-tuning (SFT) is compute efficient, it often suffers from out-of-domain (OOD) degradation. Conversely, end-to-end reinforcement learning (E2E RL) preserves OOD capabilities, but incurs high compute costs due to many turns of on-policy rollout. We introduce PivotRL, a novel framework that operates on existing SFT trajectories to combine the compute efficiency of SFT with the OOD accuracy of E2E RL. PivotRL relies on two key mechanisms: first, it executes local, on-policy rollouts and filters for pivots: informative intermediate turns where sampled actions exhibit high variance in outcomes; second, it utilizes rewards for functional-equivalent actions rather than demanding strict string matching with the SFT data demonstration. We theoretically show that these mechanisms incentivize strong learning signals with high natural gradient norm, while maximally preserving policy probability ordering on actions unrelated to training tasks. In comparison to standard SFT on identical data, we demonstrate that PivotRL achieves +4.17% higher in-domain accuracy on average across four agentic domains, and +10.04% higher OOD accuracy in non-agentic tasks. Notably, on agentic coding tasks, PivotRL achieves competitive accuracy with E2E RL with 4x fewer rollout turns. PivotRL is adopted by NVIDIA's Nemotron-3-Super-120B-A12B, acting as the workhorse in production-scale agentic post-training.

Junkeun Yi, Damon Mosk-Aoyama, Baihe Huang +9

NVIDIA · UC Berkeley

RL Post-TrainingRL for LLM AgentsAgentic RLReinforcement LearningGeneralization in RL
AskPDF
Mar 20, 2026

A Subgoal-driven Framework for Improving Long-Horizon LLM Agents

Large language model (LLM)-based agents have emerged as powerful autonomous controllers for digital environments, including mobile interfaces, operating systems, and web browsers. Web navigation, for example, requires handling dynamic content and long sequences of actions, making it particularly challenging. Existing LLM-based agents struggle with long-horizon planning in two main ways. During online execution, they often lose track as new information arrives, lacking a clear and adaptive path toward the final goal. This issue is further exacerbated during reinforcement learning (RL) fine-tuning, where sparse and delayed rewards make it difficult for agents to identify which actions lead to success, preventing them from maintaining coherent reasoning over extended tasks. To address these challenges, we propose two contributions. First, we introduce an agent framework that leverages proprietary models for online planning through subgoal decomposition. Second, we present MiRA (Milestoning your Reinforcement Learning Enhanced Agent), an RL training framework that uses dense, milestone-based reward signals. The real-time planning mechanism improves proprietary models such as Gemini by approximately a 10% absolute increase in success rate (SR) on the WebArena-Lite benchmark. Meanwhile, applying MiRA to the open Gemma3-12B model increases its success rate from 6.4% to 43.0%. This performance surpasses proprietary systems such as GPT-4-Turbo (17.6%) and GPT-4o (13.9%), as well as the previous open-model state of the art, WebRL (38.4%). Overall, our findings demonstrate that combining explicit inference-time planning with milestone-based rewards significantly improves an agent's long-horizon capabilities, paving the way for more robust and general-purpose autonomous systems.

Taiyi Wang, Sian Gooding, Florian Hartmann +2

Google DeepMind

Long-Horizon LLM AgentsWeb AgentsRL for LLM AgentsReward Shaping
AskPDF
Mar 19, 2026

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation

We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance in the 2025 International Mathematical Olympiad (IMO), the International Olympiad in Informatics (IOI), and the ICPC World Finals, demonstrating remarkably high intelligence density with 20x fewer parameters. In contrast to Nemotron-Cascade 1, the key technical advancements are as follows. After SFT on a meticulously curated dataset, we substantially expand Cascade RL to cover a much broader spectrum of reasoning and agentic domains. Furthermore, we introduce multi-domain on-policy distillation from the strongest intermediate teacher models for each domain throughout the Cascade RL process, allowing us to efficiently recover benchmark regressions and sustain strong performance gains along the way. We release the collection of model checkpoint and training data.

Zhuolin Yang, Zihan Liu, Yang Chen +14

RL for Language Model ReasoningRL for LLM AgentsReinforcement LearningLanguage Model DistillationOn-Policy Distillation
AskPDF
Mar 19, 2026

RewardFlow: Topology-Aware Reward Propagation on State Graphs for Agentic RL with Large Language Models

Reinforcement learning (RL) shows promise for enhancing LLM agentic reasoning, yet sparse terminal rewards hinder fine-grained optimization. Process reward modeling offers an alternative but incurs high computational costs, reward hacking risks, and annotation bottlenecks. We introduce RewardFlow, a lightweight method for estimating state-level rewards in agentic reasoning. By constructing state graphs that capture the intrinsic topological structure of trajectories, RewardFlow performs topology-aware propagation to estimate each state's contribution to success, yielding principled, annotation-free dense rewards. Used for RL optimization, RewardFlow substantially outperforms prior baselines across four agentic benchmarks: +6.2% average success rate on text-based tasks, +29.7% on visual reasoning over the strongest baseline across three model scales, and +10% accuracy on DeepResearch, with superior robustness and training efficiency. The implementation of RewardFlow is publicly available at https://github.com/tmlr-group/RewardFlow.

Xiao Feng, Bo Han, Zhanke Zhou +5

TMLR Group, Hong Kong Baptist University · TCL Corporate Research (HK) Co., Ltd · Cooperative Medianet Innovation Center, Shanghai Jiao Tong University +1

RL for LLM AgentsReward Design for RLDense Reward RLCredit Assignment in RL
AskPDF
Mar 19, 2026

ProRL Agent: Rollout-as-a-Service for RL Training of Multi-Turn LLM Agents

Multi-turn LLM agents are increasingly important for solving complex, interactive tasks, and reinforcement learning (RL) is a key ingredient for improving their long-horizon behavior. However, RL training requires generating large numbers of sandboxed rollout trajectories, and existing infrastructures often couple rollout orchestration with the training loop, making systems hard to migrate and maintain. Under the rollout-as-a-service philosophy, we present ProRL Agent , a scalable infrastructure that serves the full agentic rollout lifecycle through an API service. ProRL Agent also provides standardized and extensible sandbox environments that support diverse agentic tasks in rootless HPC settings. We validate ProRL Agent through RL training on software engineering, math, STEM, and coding tasks. ProRL Agent is open-sourced and integrated as part of NVIDIA NeMo Gym.

Hao Zhang, Mingjie Liu, Shaokun Zhang +10

LLM Agent TrainingLLM Agent HarnessesRL for LLM Agents
AskPDF
Mar 19, 2026

Memento-Skills: Let Agents Design Agents

We introduce Memento-Skills, a generalist, continually-learnable LLM agent system that functions as an agent-designing agent: it autonomously constructs, adapts, and improves task-specific agents through experience. The system is built on a memory-based reinforcement learning framework with stateful prompts, where reusable skills (stored as structured markdown files) serve as persistent, evolving memory. These skills encode both behaviour and context, enabling the agent to carry forward knowledge across interactions. Starting from simple elementary skills (like Web search and terminal operations), the agent continually improves via the Read--Write Reflective Learning mechanism introduced in Memento~2~\cite{wang2025memento2}. In the read phase, a behaviour-trainable skill router selects the most relevant skill conditioned on the current stateful prompt; in the write phase, the agent updates and expands its skill library based on new experience. This closed-loop design enables continual learning without updating LLM parameters, as all adaptation is realised through the evolution of externalised skills and prompts. Unlike prior approaches that rely on human-designed agents, Memento-Skills enables a generalist agent to design agents end-to-end for new tasks. Through iterative skill generation and refinement, the system progressively improves its own capabilities. Experiments on the General AI Assistants benchmark and Humanity's Last Exam demonstrate sustained gains, achieving 26.2% and 116.2% relative improvements in overall accuracy, respectively. Code is available at https://github.com/Memento-Teams/Memento-Skills.

Huichi Zhou, Siyuan Guo, Anjie Liu +14

Agentic AILLM Agent Skill LearningContinual Learning for LLM AgentsMemory-Augmented Language Model AgentsRL for LLM AgentsLLM Agent Skill Retrieval
AskPDF
Mar 19, 2026

HISR: Hindsight Information Modulated Segmental Process Rewards For Multi-turn Agentic Reinforcement Learning

While large language models excel in diverse domains, their performance on complex longhorizon agentic decision-making tasks remains limited. Most existing methods concentrate on designing effective reward models (RMs) to advance performance via multi-turn reinforcement learning. However, they suffer from delayed propagation in sparse outcome rewards and unreliable credit assignment with potentially overly fine-grained and unfocused turnlevel process rewards. In this paper, we propose (HISR) exploiting Hindsight Information to modulate Segmental process Rewards, which closely aligns rewards with sub-goals and underscores significant segments to enhance the reliability of credit assignment. Specifically, a segment-level process RM is presented to assign rewards for each sub-goal in the task, avoiding excessively granular allocation to turns. To emphasize significant segments in the trajectory, a hindsight model is devised to reflect the preference of performing a certain action after knowing the trajectory outcome. With this characteristic, we design the ratios of sequence likelihoods between hindsight and policy model to measure action importance. The ratios are subsequently employed to aggregate segment importance scores, which in turn modulate segmental process rewards, enhancing credit assignment reliability. Extensive experimental results on three publicly benchmarks demonstrate the validity of our method.

Zhicong Lu, Zichuan Lin, Wei Jia +7

Aerospace Information Research Institute, Chinese Academy of Sciences · University of Chinese Academy of Sciences · Tencent Hunyuan +1

RL for LLM AgentsAgentic RLProcess Reward ModelsCredit Assignment in RLSparse-Reward RL
AskPDF
Mar 18, 2026

CodeScout: An Effective Recipe for Reinforcement Learning of Code Search Agents

A prerequisite for coding agents to perform tasks on large repositories is code localization - the identification of relevant files, classes, and functions to work on. While repository-level code localization has been performed using embedding-based retrieval approaches such as vector search, recent work has focused on developing agents to localize relevant code either as a standalone precursor to or interleaved with performing actual work. Most prior methods on agentic code search equip the agent with complex, specialized tools, such as repository graphs derived from static analysis. In this paper, we demonstrate that, with an effective reinforcement learning recipe, a coding agent equipped with nothing more than a standard Unix terminal can be trained to achieve strong results. Our experiments on three benchmarks (SWE-Bench Verified, Pro, and Lite) reveal that our models consistently achieve superior or competitive performance over 2-18x larger base and post-trained LLMs and sometimes approach performance provided by closed models like Claude Sonnet, even when using specialized scaffolds. Our work particularly focuses on techniques for re-purposing existing coding agent environments for code search, reward design, and RL optimization. We release the resulting model family, CodeScout, along with all our code and data for the community to build upon.

Lintang Sutawika, Aditya Bharat Soni, Bharath Sriraam R R +8

Carnegie Mellon University · Independent · OpenHands

Code LocalizationCode SearchRL for LLM AgentsReinforcement LearningInformation Retrieval
AskPDF
Mar 18, 2026

Towards Reliable Local Security Agents: Verifiable Post-Training for Linux Privilege Escalation

LLM agents are becoming increasingly important in the security domain, but leading systems are often closed-source, cloud-based, hard to reproduce or use with sensitive code. This creates a need for small, local models that can perform security tasks under strict resource constraints, though effective methods for developing them remain unexplored. In this paper, we address this gap by proposing a two-stage post-training recipe that turns a small local language model into a security agent. To this end, we focus on Linux privilege escalation as a representative setting to systematically study the training of local models, as the task is both automatically verifiable and requires multi-step interactive reasoning. Using an experimental setup that mitigates data leakage, we post-train a small 4B model in two stages: supervised fine-tuning on traces from procedural privilege-escalation environments, followed by reinforcement learning with verifiable rewards. On a held-out benchmark of 12 Linux privilege-escalation scenarios, supervised fine-tuning doubles the baseline success rate under a tight budget of 20 interaction rounds, and subsequent reinforcement learning training improves our model, PrivEsc-LLM 4B, to 93.3% success, behind only Claude Opus 4.7 at this budget. At the same time, the expected inference cost per successful escalation decreases by more than 80x. Our findings not only show that small local models can be adapted to complex security tasks, but also document the challenges involved, offering guidance for transferring this recipe to other settings.

Philipp Normann, Andreas Happe, Jürgen Cito +1

TU Wien, Vienna, Austria

LLM Agent SecurityLLM Post-TrainingRL for LLM Agents
AskPDF
Mar 18, 2026

Complementary RL: Towards Efficient Experience-Driven Agent Learning

Reinforcement Learning (RL) has emerged as a powerful paradigm for training LLM-based agents, yet remains limited by low sample efficiency, stemming not only from sparse outcome feedback but also from the agent's inability to leverage prior experience across episodes. While augmenting agents with historical experience offers a promising remedy, existing approaches suffer from a critical weakness: the experience distilled from history is either stored statically or fail to coevolve with the improving actor, causing a progressive misalignment between the experience and the actor's evolving capability that diminishes its utility over the course of training. Inspired by complementary learning systems in neuroscience, we present Complementary RL to achieve seamless co-evolution of an experience extractor and a policy actor within the RL optimization loop. Specifically, the actor is optimized via sparse outcome-based rewards, while the experience extractor is optimized according to whether its distilled experiences demonstrably contribute to the actor's success, thereby evolving its experience management strategy in lockstep with the actor's growing capabilities. Empirically, Complementary RL outperforms outcome-based agentic RL baselines that do not learn from experience, achieving 10% performance improvement in single-task scenarios and exhibits robust scalability in multi-task settings. These results establish Complementary RL as a paradigm for efficient experience-driven agent learning.

Dilxat Muhtar, Jiashun Liu, Wei Gao +8

Alibaba Group · HKUST

Reinforcement LearningSample-Efficient RLRL for LLM Agents
AskPDF
Mar 18, 2026

SLEA-RL: Step-Level Experience Augmented Reinforcement Learning for Multi-Turn Agentic Training

Large Language Model (LLM) agents have shown strong results on multi-turn tool-use tasks, yet they operate in isolation during training, failing to leverage experiences accumulated across episodes. Existing experience-augmented methods address this by organizing trajectories into retrievable libraries, but they retrieve experiences only once based on the initial task description and hold them constant throughout the episode. In multi-turn settings where observations change at every step, this static retrieval becomes increasingly mismatched as episodes progress. We propose SLEA-RL (Step-Level Experience-Augmented Reinforcement Learning), a framework that retrieves relevant experiences at each decision step conditioned on the current observation. SLEA-RL operates through three components: (i) step-level observation clustering that groups structurally equivalent environmental states for efficient cluster-indexed retrieval; (ii) a self-evolving experience library that distills successful strategies and failure patterns through score-based admission and rate-limited extraction; and (iii) policy optimization with step-level credit assignment for fine-grained advantage estimation across multi-turn episodes. The experience library evolves alongside the policy through semantic analysis rather than gradient updates. Experiments on long-horizon multi-turn agent benchmarks demonstrate that SLEA-RL achieves superior performance compared to various reinforcement learning baselines.

Prince Zizhuang Wang, Shuli Jiang

Carnegie Mellon University

RL for LLM AgentsAgentic RLReinforcement LearningMulti-Turn RLRL for Tool UseMemory-Augmented Language Model Agents
AskPDF
Mar 17, 2026

Learning to Present: Inverse Specification Rewards for Agentic Slide Generation

Automated presentation generation remains a challenging task requiring coherent content creation, visual design, and audience-aware communication. This work proposes an OpenEnv-compatible reinforcement learning environment where LLM agents learn to research topics, plan content, and generate professional HTML slide presentations through tool use. We introduce a multi-component reward system combining structural validation, render quality assessment, LLM-based aesthetic scoring, content quality metrics, and an inverse specification reward that measures how faithfully generated slides convey their intended purpose. The inverse specification reward, an "inverse task" where an LLM attempts to recover the original specification from generated slides, provides a holistic quality signal. Our approach fine-tunes Qwen2.5-Coder-7B via GRPO, training only 0.5% of parameters on prompts derived from expert demonstrations collected using Claude Opus 4.6. Experiments on 48 diverse business briefs across six models demonstrate that our fine-tuned 7B model achieves 91.2% of Claude Opus 4.6's quality while improving 33.1% over the base model. The six-model comparison reveals that instruction adherence and tool-use compliance, rather than raw parameter count, determine agentic task performance. We contribute SlideRL, an open-source dataset of 288 multi-turn rollout trajectories across all six models: https://huggingface.co/datasets/KarthikRagunathAnandaKumar/sliderl-multi-turn-rollouts Code: https://github.com/pushing-the-frontier/slide-forge-llm

Karthik Ragunath Ananda Kumar, Subrahmanyam Arunachalam

Tavus Inc. · University of Texas at Dallas · Texas A&M University

LLM Agent TrainingTool-Augmented Language Model AgentsRL for LLM Agents
AskPDF
Mar 17, 2026

When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making

Embodied robotic systems increasingly rely on large language model (LLM)-based agents to support high-level reasoning, planning, and decision-making during interactions with the environment. However, invoking LLM reasoning introduces substantial computational latency and resource overhead, which can interrupt action execution and reduce system reliability. Excessive reasoning may delay actions, while insufficient reasoning often leads to incorrect decisions and task failures. This raises a fundamental question for embodied agents: when should the agent reason, and when should it act? In this work, we propose RARRL (Resource-Aware Reasoning via Reinforcement Learning), a hierarchical framework for resource-aware orchestration of embodied agents. Rather than learning low-level control policies, RARRL learns a high-level orchestration policy that operates at the agent's decision-making layer. This policy enables the agent to adaptively determine whether to invoke reasoning, which reasoning role to employ, and how much computational budget to allocate based on current observations, execution history, and remaining resources. Extensive experiments, including evaluations with empirical latency profiles derived from the ALFRED benchmark, show that RARRL consistently improves task success rates while reducing execution latency and enhancing robustness compared with fixed or heuristic reasoning strategies. These results demonstrate that adaptive reasoning control is essential for building reliable and efficient embodied robotic agents.

Jun Liu, Pu Zhao, Zhenglun Kong +12

Robotics Institute, Carnegie Mellon University · Northeastern University · Harvard University +9

Embodied AIEmbodied ReasoningRL for LLM AgentsRobotic RL
AskPDF
Mar 17, 2026

SQL-ASTRA: Alleviating Sparse Feedback in Agentic SQL via Column-Set Matching and Trajectory Aggregation

Agentic Reinforcement Learning (RL) shows promise for complex tasks, but Text-to-SQL remains mostly restricted to single-turn paradigms. A primary bottleneck is the credit assignment problem. In traditional paradigms, rewards are determined solely by the final-turn feedback, which ignores the intermediate process and leads to ambiguous credit evaluation. To address this, we propose Agentic SQL, a framework featuring a universal two-tiered reward mechanism designed to provide effective trajectory-level evaluation and dense step-level signals. First, we introduce Aggregated Trajectory Reward (ATR) to resolve multi-turn credit assignment. Using an asymmetric transition matrix, ATR aggregates process-oriented scores to incentivize continuous improvement. Leveraging Lyapunov stability theory, we prove ATR acts as an energy dissipation operator, guaranteeing a cycle-free policy and monotonic convergence. Second, Column-Set Matching Reward (CSMR) provides immediate step-level rewards to mitigate sparsity. By executing queries at each turn, CSMR converts binary (0/1) feedback into dense [0, 1] signals based on partial correctness. Evaluations on BIRD show a 5% gain over binary-reward GRPO. Notably, our approach outperforms SOTA Arctic-Text2SQL-R1-7B on BIRD and Spider 2.0 using identical models, propelling Text-to-SQL toward a robust multi-turn agent paradigm.

Long Li, Zhijian Zhou, Jiangxuan Long +5

Griffith University, Brisbane, Australia · Fudan University, Shanghai, China · Shanghai Innovation Institute, China +4

Text-to-SQLStep-Level Credit Assignment in RLRL for LLM AgentsSparse-Reward RLAgentic RLDense Reward RL
AskPDF
Mar 17, 2026

SWE-QA-Pro: A Representative Benchmark and Scalable Training Recipe for Repository-Level Code Understanding

Agentic repository-level code understanding is essential for automating complex software engineering tasks, yet the field lacks reliable benchmarks. Existing evaluations often overlook the long tail topics and rely on popular repositories where Large Language Models (LLMs) can cheat via memorized knowledge. To address this, we introduce SWE-QA-Pro, a benchmark constructed from diverse, long-tail repositories with executable environments. We enforce topical balance via issue-driven clustering to cover under-represented task types and apply a rigorous difficulty calibration process: questions solvable by direct-answer baselines are filtered out. This results in a dataset where agentic workflows significantly outperform direct answering (e.g., a ~13-point gap for Claude Sonnet 4.5), confirming the necessity of agentic codebase exploration. Furthermore, to tackle the scarcity of training data for such complex behaviors, we propose a scalable synthetic data pipeline that powers a two-stage training recipe: Supervised Fine-Tuning (SFT) followed by Reinforcement Learning from AI Feedback (RLAIF). This approach allows small open models to learn efficient tool usage and reasoning. Empirically, a Qwen3-8B model trained with our recipe surpasses GPT-4o by 2.3 points on SWE-QA-Pro and substantially narrows the gap to state-of-the-art proprietary models, demonstrating both the validity of our evaluation and the effectiveness of our agentic training workflow.

Songcheng Cai, Zhiheng Lyu, Yuansheng Ni +13

University of Waterloo · University of Toronto · The Hong Kong University of Science and Technology +2

AI Agent BenchmarksSoftware Engineering BenchmarksSoftware EngineeringLLM Agent TrainingRL for LLM Agents
AskPDF
Mar 13, 2026

Joint Optimization of Multi-agent Memory System

Memory systems are critical for LLMs, mitigating context window limitations and supporting long-horizon user-LLM interactions. Such systems typically comprise multiple agents responsible for memory construction and retrieval. Existing approaches often optimize each agent independently under a shared global objective (e.g., downstream QA accuracy), treating other agents as a static environment. However, this design has two key limitations: (1) independent optimization ignores inter-agent dependencies and lacks agents' co-adaptation, and (2) relying solely on sparse global rewards provides limited guidance for optimizing specialized agents and causes ambiguous credit assignment. These may ultimately limit agents' effective collaboration in the memory system. To address these limitations, we propose CoMAM, a joint optimization framework that promotes collaboration among agents via end-to-end reinforcement learning and an adaptive credit assignment mechanism. Specifically, we model the multi-agent pipeline as a Markov decision process (MDP) to expose inter-agent dependencies during end-to-end training. Agents are then jointly optimized using a combination of their local task reward and an adaptively weighted global reward, enabling agents to co-adapt while receiving targeted feedback for their respective roles. Experiments show that CoMAM consistently outperforms leading memory systems, validating the effectiveness of the joint optimization framework.

Wenyu Mao, Haoyang Liu, Haosong Tan +4

University of Science and Technology of China · Institute of Dataspace, Hefei Comprehensive National Science Center

Memory-Augmented Language Model AgentsMulti-Agent System OptimizationMulti-Agent CoordinationRL for LLM AgentsMulti-Agent Reinforcement Learning
AskPDF
Mar 12, 2026

On Information Self-Locking in Reinforcement Learning for Active Reasoning of LLM agents

Reinforcement learning (RL) has become a de facto paradigm for building LLM-based agents that act, interact, and reason over extended task horizons. However, in active reasoning where agents must elicit new observations through interaction with the environment to solve the task, we find that outcome-based RL can induce a systematic failure mode which we call information self-locking (SeL): agents fail both to elicit informative feedback and to internalize obtained evidence. To understand the issue, we trace agentic behaviors into two coupled capabilities: Action Selection (AS), which determines observation streams, and Belief Tracking (BT), which updates the agent's internal task understanding. Theoretical and empirical analyses reveal a bidirectional bottleneck that leads to SeL: weak BT obscures the credit of informative actions, while weak AS deprives BT of useful evidence. This coupling weakens the learning signal for both capabilities and leads to SeL. To mitigate this issue, we propose AREW, a simple yet effective Advantage Reweighting method that uses easy-to-obtain directional critiques to reallocate credit within trajectories. Extensive experiments across 9 agentic tasks of varying complexity show that AREW significantly mitigates SeL, yielding up to 60-point gains in final performance. Code is available at https://github.com/unimpor/T3.

Deyu Zou, Yongqiang Chen, Fan Feng +4

The Chinese University of Hong Kong · University of California San Diego · Georgia Institute of Technology +1

RL for LLM AgentsActive Information GatheringStep-Level Credit Assignment in RL
AskPDF
Mar 12, 2026

Can RL Improve Generalization of LLM Agents? An Empirical Study

Reinforcement fine-tuning (RFT) has shown promise for training LLM agents to perform multi-turn decision-making based on environment feedback. However, most existing evaluations remain largely in-domain: training and testing are conducted in the same environment or even on the same tasks. In real-world deployment, agents may operate in unseen environments with different background knowledge, observation spaces, and action interfaces. To characterize the generalization profile of RFT under such shifts, we conduct a systematic study along three axes: (1) within-environment generalization across task difficulty, (2) cross-environment transfer to unseen environments, and (3) sequential multi-environment training to quantify transfer and forgetting. Our results show that RFT generalizes well across task difficulty within an environment, but exhibits weaker transfer to unseen environments, which correlates with shifts in both semantic priors and observation/action interfaces. In contrast, sequential training yields promising downstream gains with minimal upstream forgetting, and mixture training across environments improves the overall balance. We further provide detailed analyses and deeper insights, and hope our work helps the community develop and deploy generalizable LLM agents.

Zhiheng Xi, Xin Guo, Jiaqi Liu +11

Fudan University · Shanghai Artificial Intelligence Laboratory · Meituan

LLM AgentsRL for LLM AgentsRL Fine-TuningGeneralization in RLContinual Learning for LLM AgentsMulti-Task RL
AskPDF
Mar 10, 2026

Evaluate-as-Action: Self-Evaluated Process Rewards for Retrieval-Augmented Agents

Retrieval-augmented agents can query external evidence, yet their reliability in multi-step reasoning remains limited: noisy retrieval may derail multi-hop question answering, while outcome-only reinforcement learning provides credit signals that are too coarse to optimize intermediate steps. We propose EvalAct (Evaluate-as-Action), which converts implicit retrieval quality assessment into an explicit action and enforces a coupled Search-to-Evaluate protocol so that each retrieval is immediately followed by a structured evaluation score, yielding process signals aligned with the interaction trajectory. To leverage these signals, we introduce Process-Calibrated Advantage Rescaling (PCAR), a GRPO-based optimization method that rescales advantages at the segment level according to evaluation scores, emphasizing reliable segments while updating uncertain ones conservatively. Experiments on seven open-domain QA benchmarks show that EvalAct achieves the best average accuracy, with the largest gains on multi-hop tasks, and ablations verify that the explicit evaluation loop drives the primary improvements while PCAR provides consistent additional benefits.

Jiangming Shu, Yuxiang Zhang, Ye Ma +2

School of Computer Science and Technology, Beijing Jiaotong University · Hithink Research

Retrieval-Augmented ReasoningAgentic RAGRL for LLM AgentsProcess Reward ModelsMulti-Hop QAGenerative AI Evaluation
AskPDF