ECHO is a clipped policy-gradient method that uses posterior-sensitive rewards to give turn-level epistemic credit in multi-turn information-seeking tasks, outperforming trajectory-level GRPO on a new Clue Selector Game benchmark.
hub Canonical reference
Process reward models for llm agents: Practical framework and directions
Canonical reference. 83% of citing Pith papers cite this work as background.
16
Pith papers citing it
Background
83% of classified citations
hub tools
citation-role summary
background 6
citation-polarity summary
roles
background 6representative citing papers
TRACER attaches verifiable sentence-level provenance records to multimodal agent outputs using tool-turn alignment and semantic relations, yielding 78.23% answer accuracy and fewer tool calls than baselines on TRACE-Bench.
The log-probability ratio from RL post-training recovers the optimal advantage function, providing an effective free signal for test-time scaling, uncertainty estimation, and failure attribution in LLM agents.
SCPO recovers step-level credit from successful siblings within rollout groups to reduce semantic inconsistency in group-based RL for LLM agents, matching or exceeding baselines on ALFWorld and WebShop.
Success Visitation Matching uses a discriminator to turn sparse outcome rewards into dense process rewards by matching visitations of successful episodes, provably preserving the optimal policy and speeding up robotic RL finetuning.
Action-conditioned estimation of intervention advantage via prefix branching reduces control regret over calibrated scalar risk scores in LLM agent oversight across benchmarks.
ARCO co-trains an LLM agent with a per-step rubric model whose natural-language criteria and scoring function evolve with the policy, and reports the best exact-match accuracy in all six benchmark/backbone settings compared.
METIS internalizes curriculum judgment in LLM reinforcement fine-tuning by predicting within-prompt reward variance via in-context learning and jointly optimizing with a self-judgment reward, yielding superior performance and up to 67% faster convergence across math, code, and agent benchmarks.
A²TGPO improves RL policy optimization for multi-turn agentic LLMs by normalizing information gain within same-depth turn groups, rescaling cumulative advantages by sqrt of term count, and modulating clipping ranges per turn's normalized IG.
ReRec uses reinforcement fine-tuning with dual-graph reward shaping, reasoning-aware advantage estimation, and online curriculum scheduling to improve LLM reasoning and performance in recommendation tasks.
Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.
Q-Evolve unifies automatic process-reward labeling via advantage estimation and behavior-proximal policy optimization inside an in-distribution RL loop to enable self-evolving LLM agents on interactive tasks.
StainFlow proposes global entity stain tracking and local stain evidence linking modules to improve process rewards for GUI agents, reporting 3.2% relative gain in online RL success and 1.8% in judgment accuracy on AndroidWorld and OGRBench.
SMCS coordinates 15 open-source LLMs via retrieval-based prior selection and exploration-exploitation posterior enhancement, outperforming GPT-4.1 by 5.36% and GPT-o3-mini by 5.28% on eight benchmarks.
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
The paper delivers the first systematic review of self-evolving agents, structured around what components evolve, when adaptation occurs, and how it is implemented.
citing papers explorer
-
ECHO: Learning Epistemically Adaptive Language Agents with Turn-Level Credit
ECHO is a clipped policy-gradient method that uses posterior-sensitive rewards to give turn-level epistemic credit in multi-turn information-seeking tasks, outperforming trajectory-level GRPO on a new Clue Selector Game benchmark.
-
TRACER: Verifiable Generative Provenance for Multimodal Tool-Using Agents
TRACER attaches verifiable sentence-level provenance records to multimodal agent outputs using tool-turn alignment and semantic relations, yielding 78.23% answer accuracy and fewer tool calls than baselines on TRACE-Bench.
-
Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents
The log-probability ratio from RL post-training recovers the optimal advantage function, providing an effective free signal for test-time scaling, uncertainty estimation, and failure attribution in LLM agents.
-
Semantic Consistency Policy Optimization for Reinforcement Learning of LLM Agents
SCPO recovers step-level credit from successful siblings within rollout groups to reduce semantic inconsistency in group-based RL for LLM agents, matching or exceeding baselines on ALFWorld and WebShop.
-
Learning Process Rewards via Success Visitation Matching for Efficient RL
Success Visitation Matching uses a discriminator to turn sparse outcome rewards into dense process rewards by matching visitations of successful episodes, provably preserving the optimal policy and speeding up robotic RL finetuning.
-
Calibration Is Not Control: Why LLM-Agent Oversight Needs Intervention
Action-conditioned estimation of intervention advantage via prefix branching reduces control regret over calibrated scalar risk scores in LLM agent oversight across benchmarks.
-
ARCO: Adaptive Rubrics with Co-Evolution for Multi-Step LLM-Based Agents
ARCO co-trains an LLM agent with a per-step rubric model whose natural-language criteria and scoring function evolve with the policy, and reports the best exact-match accuracy in all six benchmark/backbone settings compared.
-
Internalizing Curriculum Judgment for LLM Reinforcement Fine-Tuning
METIS internalizes curriculum judgment in LLM reinforcement fine-tuning by predicting within-prompt reward variance via in-context learning and jointly optimizing with a self-judgment reward, yielding superior performance and up to 67% faster convergence across math, code, and agent benchmarks.
-
A$^2$TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level Clipping
A²TGPO improves RL policy optimization for multi-turn agentic LLMs by normalizing information gain within same-depth turn groups, rescaling cumulative advantages by sqrt of term count, and modulating clipping ranges per turn's normalized IG.
-
ReRec: Reasoning-Augmented LLM-based Recommendation Assistant via Reinforcement Fine-tuning
ReRec uses reinforcement fine-tuning with dual-graph reward shaping, reasoning-aware advantage estimation, and online curriculum scheduling to improve LLM reasoning and performance in recommendation tasks.
-
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.
-
Self-evolving LLM agents with in-distribution Optimization
Q-Evolve unifies automatic process-reward labeling via advantage estimation and behavior-proximal policy optimization inside an in-distribution RL loop to enable self-evolving LLM agents on interactive tasks.
-
StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents
StainFlow proposes global entity stain tracking and local stain evidence linking modules to improve process rewards for GUI agents, reporting 3.2% relative gain in online RL success and 1.8% in judgment accuracy on AndroidWorld and OGRBench.
-
A Scalable Multi-LLM Collaboration System with Retrieval-based Selection and Exploration-Exploitation-Driven Enhancement
SMCS coordinates 15 open-source LLMs via retrieval-based prior selection and exploration-exploitation posterior enhancement, outperforming GPT-4.1 by 5.36% and GPT-o3-mini by 5.28% on eight benchmarks.
-
Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
-
A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence
The paper delivers the first systematic review of self-evolving agents, structured around what components evolve, when adaptation occurs, and how it is implemented.