SPADER is a novel reinforcement learning framework that addresses the challenges of Multi-Answer Question Answering by improving credit assignment and promoting diverse exploration during long-horizon…

The paper proposes a novel temporal and structural credit assignment framework to efficiently optimize multi-agent LLM systems by decomposing the error signal and using targeted, discrete gradient upd…

The paper proposes RoRo, a rubric-guided process reward framework that improves stepwise model routing by evaluating the quality of intermediate reasoning steps, leading to better performance and cost…

LongTraceRL addresses long-context reasoning challenges by generating highly challenging training data and introducing a fine-grained rubric reward, significantly improving evidence-grounded reasoning…

The paper proposes Skill-RM, a unified framework that treats reward modeling as an agentic task to consistently integrate diverse evaluation criteria, achieving superior performance over traditional m…

The paper proposes Deep Research as Rubric (DR-rubric), a novel evidence-driven framework that treats rubric construction itself as a research problem to generate fine-grained, scalable reward signals…

The paper introduces Latent Reward Steering (LRS), an adaptive inference-time framework that implicitly improves the reasoning ability of LLMs by guiding the model's internal latent states based on a…

The paper proposes MERIT, a dual-level, multi-horizon memory retrieval framework that significantly improves the performance of interactive text-to-SQL agents by providing both global and local memory…

TRACER introduces a novel turn-level reinforcement framework that enables cooperative multi-LLM reasoning by separating decision-making into a regret-matching controller and a generation-credit layer.

The paper introduces an outer-loop AI agent that autonomously redesigns LLM policy-synthesis pipelines for multi-agent social dilemmas, demonstrating that the optimal pipeline structure depends critic…

This paper introduces RREDCoT, a method for approximating optimal reward redistribution in Chain-of-Thought reasoning language models without additional generation.

This paper introduces RREDCoT, a method for approximating optimal reward redistribution in Chain-of-Thought reasoning language models without additional generation.

The paper proposes DecomposeR, a planner-centric framework that structures deep research into typed Directed Acyclic Graphs (DAGs) to explicitly improve the planning and execution of large language mo…

PRO-CUA introduces a process-reward optimization framework that enables efficient, step-level reinforcement learning for training computer use agents by decoupling environment interaction from policy…

The paper proposes ProRL, an effective Reinforcement Learning framework that rectifies gradient estimation deficiencies to optimize proactive recommendation paths, significantly outperforming existing…

The paper proposes Credit-Attenuated Privileged Feedback (CAPF), a training-time mechanism that uses verifier-side information to guide LLM search agents, significantly improving their performance on…

ExpGraph is a model-agnostic framework that uses a self-evolving experience graph to enable LLM agents to reuse past successful strategies and failure lessons, significantly improving performance acro…

The paper introduces Plan, a structured agentic behavior that decomposes multi-hop questions into ordered sub-questions before retrieval, and proposes a self-bootstrapping paradigm to train it without…

The paper introduces Expected Value Alignment (EVA), a novel reward modeling procedure that allows continuous scoring of intermediate reasoning steps in formal mathematics verification while maintaini…

The paper introduces VibeSearchBench, a new benchmark designed to evaluate long-horizon, proactive search capabilities, demonstrating that current state-of-the-art LLM agents are still significantly i…