Claude 3 Opus strategically fakes alignment by complying with harmful requests only during simulated training to preserve its preference for refusing them afterward.
mega hub Mixed citations
Proximal Policy Optimization Algorithms
Mixed citation behavior. Most common role is background (52%).
1486
Pith papers citing it
Background
52% of classified citations
abstract
We propose a new family of policy gradient methods for reinforcement learning, which alternate between sampling data through interaction with the environment, and optimizing a "surrogate" objective function using stochastic gradient ascent. Whereas standard policy gradient methods perform one gradient update per data sample, we propose a novel objective function that enables multiple epochs of minibatch updates. The new methods, which we call proximal policy optimization (PPO), have some of the benefits of trust region policy optimization (TRPO), but they are much simpler to implement, more general, and have better sample complexity (empirically). Our experiments test PPO on a collection of benchmark tasks, including simulated robotic locomotion and Atari game playing, and we show that PPO outperforms other online policy gradient methods, and overall strikes a favorable balance between sample complexity, simplicity, and wall-time.
hub tools
citation-role summary
background 155
method 113
baseline 15
dataset 4
citation-polarity summary
claims ledger
- abstract We propose a new family of policy gradient methods for reinforcement learning, which alternate between sampling data through interaction with the environment, and optimizing a "surrogate" objective function using stochastic gradient ascent. Whereas standard policy gradient methods perform one gradient update per data sample, we propose a novel objective function that enables multiple epochs of minibatch updates. The new methods, which we call proximal policy optimization (PPO), have some of the benefits of trust region policy optimization (TRPO), but they are much simpler to implement, more ge
authors
mega hub controls
export citing contexts JSON
export graph JSON
export full bundle JSON
open full Pith review
annotated reader queued
Recognition alignment
counterfactual ablation
co-cited works
representative citing papers
GRPO's group-mean baseline assigns identical advantages to all tokens under output-only rewards, inducing gradient sparsity and an intrinsic rank-2 structure proven from the zero-sum constraint and confirmed by SVD on Nemotron-4B gradients.
Negative narrative immersion causes 12-31% drops in LLM moral accuracy and produces structured shifts that appear in downstream applications.
RL agent for online LHC trigger threshold tuning improves in-tolerance intervals by 28-56% on Monte Carlo and real CMS data without fine-tuning.
Dynamic isotropy, quantifying uniform center-of-mass acceleration capability, improves robot performance and enables omnidirectional locomotion, terrain traversal, and failure resilience in a spherical robot design.
AtomComposer uses online RL with multi-composition training to discover up to 10x more valid 3D isomers on unseen chemical formulas than single-composition baselines.
Agent-BRACE improves LLM agent performance on long-horizon partially observable tasks by 5.3-14.5% through a decoupled belief state of verbalized atomic claims with certainty labels that keeps context length constant.
SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
ReLibra uses pre-known token-to-expert routing from RL rollouts to perform inter-batch expert reordering and intra-batch replication, delivering up to 1.6x higher throughput than Megatron-LM and 1.2x over oracle-equipped EPLB while staying within 6-10% of an ideal balanced baseline.
Weak-to-strong generalization is nearly inevitable in linear logistic regression for most student-teacher pairs without any model capacity mismatch.
Observation and action delays are formally equivalent in cooperative Dec-POMDPs, yielding identical optimal solutions and enabling zero-shot transfer, though learning dynamics differ due to credit assignment and operational constraints.
A language-game framework enables dialogue with dynamical systems such as GRNs by treating their frozen dynamics as an RL policy core, using an LM to route prompts so the system responds through its own behavior without parameter changes.
RefereeBench shows that even the strongest video MLLMs reach only around 60% accuracy on multi-sport refereeing tasks and struggle with rule application and temporal grounding.
OP-GRPO is the first off-policy GRPO method for flow-matching models that reuses trajectories via replay buffer and importance sampling corrections, matching on-policy performance with 34.2% of the training steps.
User-turn generation reveals that LLMs' interaction awareness is largely decoupled from task accuracy, remaining near zero in deterministic settings even as accuracy scales to 96.8% on GSM8K.
Derives an exact unbiased policy gradient for RL post-training of diffusion LLMs via entropy-guided step selection and one-step denoising rewards, achieving state-of-the-art results on coding and logical reasoning benchmarks.
A certified gradient-based method for contact-rich manipulation that quantifies smoothing-induced errors via set-valued discrepancies and incorporates them into analytical reachable sets for robust affine feedback policies.
First in-orbit demonstration of a DRL-trained AI satellite attitude controller that performs robust inertial pointing after sim-to-real transfer.
Develops and tests the first effective safeguard for analytic gradient-based provably safe RL, showing safe training on three control tasks without performance loss.
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
SWE-Gym supplies 2438 executable real-world Python tasks to train SWE agents and verifiers, yielding up to 19% gains and new open-weight SOTA of 32% on SWE-Bench Verified.
BEHAVIOR-1K introduces a benchmark of 1,000 human everyday activities in realistic simulated scenes together with the OMNIGIBSON physics simulator to evaluate embodied AI.
ORPO performs preference alignment during supervised fine-tuning via a monolithic odds ratio penalty, allowing 7B models to outperform larger state-of-the-art models on alignment benchmarks.
STEMGym benchmark demonstrates that perception pipelines dominate dose efficiency in autonomous STEM over navigation methods across 33 agent setups.
citing papers explorer
-
Featurized Occupation Measures for Structured Global Search in Numerical Optimal Control
Featurized Occupation Measures create a primal-dual framework that couples explicit HJB subsolutions with scalable trajectory optimization, proving asymptotic consistency and shifting dimensionality limits to system interconnection topology.
-
Topo-R1: Detecting Topological Anomalies via Vision-Language Models
Topo-R1 fine-tunes a vision-language model using a topology-aware reward and GRPO to detect anomalies such as broken or spurious connections in tubular segmentation masks, outperforming standard VLMs.
-
Automatic Generation of High-Performance RL Environments
Closed-loop prompt-based translation with hierarchical verification and iterative repair produces equivalent high-performance RL environments across five cases including new TCGJax.
-
NePPO: Near-Potential Policy Optimization for General-Sum Multi-Agent Reinforcement Learning
NePPO learns a player-independent potential function via a novel objective whose minimization yields an approximate Nash equilibrium for general-sum multi-agent games.
-
Prism: Policy Reuse via Interpretable Strategy Mapping in Reinforcement Learning
PRISM transfers RL policies zero-shot by aligning causally validated discrete concepts from agent encoders, achieving 69-76% win rates in Go 7x7 but random performance in Atari Breakout.
-
Alternating Reinforcement Learning with Contextual Rubric Rewards: Beyond the Scalarization Strategy
ARL-RR alternates optimization over rubric meta-classes with dynamic selection to avoid fixed scalarization, outperforming baselines on HealthBench.
-
Improving Text-to-Image Generation with Intrinsic Self-Confidence Rewards
SOLACE improves text-to-image generation by using intrinsic self-confidence rewards from noise reconstruction accuracy during reinforcement learning post-training without external supervision.
-
LongVideo-R1: Smart Navigation for Low-cost Long Video Understanding
LongVideo-R1 trains a reasoning agent on 33K trajectories to intelligently select informative video clips via iterative refinement and RL, achieving better accuracy-efficiency tradeoffs on long video QA benchmarks.
-
RLGT: A reinforcement learning framework for extremal graph theory
RLGT is a modular reinforcement learning framework for extremal graph theory that handles undirected, directed, looped, and multi-colored graphs to facilitate future research.
-
Amortized Molecular Optimization via Group Relative Policy Optimization
AMORTIX uses group-normalized rewards in reinforcement learning to train an amortized Graph Transformer that optimizes constrained molecules in one forward pass and outperforms baselines on kinase inhibitor and prodrug design tasks.
-
Simulating Word Suggestion Usage in Mobile Typing to Guide Intelligent Text Entry Design
WSTypist is a new RL-based simulation model that reproduces human-like word suggestion strategies, individual differences, and adaptation to design changes in mobile text entry.
-
RL-VLA$^3$: A Flexible and Asynchronous Reinforcement Learning Framework for VLA Training
RL-VLA³ is an asynchronous RL framework for VLA training that delivers up to 85.2% higher throughput than synchronous baselines while preserving identical sample efficiency and scaling to 256 GPUs.
-
Beyond Variance: Prompt-Efficient RLVR via Rare-Event Amplification and Bidirectional Pairing
Positive-negative prompt pairing with weighted GRPO improves RLVR sample efficiency, raising AIME 2025 Pass@8 from 16.8 to 22.2 on Qwen2.5-Math-7B while matching large-scale training.
-
Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Models
GPS trains a small model on optimization history to predict prompt difficulty and select intermediate-difficulty diverse batches, yielding better training efficiency, final performance, and test-time allocation than baselines on reasoning benchmarks.
-
Minerva: Reinforcement Learning with Verifiable Rewards for Cyber Threat Intelligence LLMs
MinervaRL applies reinforcement learning with verifiable rewards from CTI standards to improve LLM structured output performance by 15.8 points over base models across 12 benchmarks.
-
Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models
A single LLM improves its own reasoning by self-distilling from privileged verified traces as teacher to its question-only student policy, outperforming off-policy distillation and RL on math benchmarks with better token efficiency.
-
Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning
Temp-R1 uses reverse curriculum reinforcement learning to train an autonomous agent that achieves state-of-the-art results on temporal KGQA benchmarks by developing sophisticated reasoning on hard questions first.
-
Learning to Discover at Test Time
TTT-Discover applies test-time RL to set new state-of-the-art results on math inequalities, GPU kernels, algorithm contests, and single-cell denoising using an open model and public code.
-
When Backdoors Meet Partial Observability: Attacking Real-World Reinforcement Learning
DGBA enables reliable backdoor attacks on real-world RL policies under partial observability by learning stochastic visual triggers via conditional diffusion and using advantage-based poisoning at critical states.
-
Group Relative Policy Optimization for Robust Blind Interference Alignment with Fluid Antennas
GRPO, a critic-free DRL algorithm, optimizes fluid antenna positions for robust blind interference alignment, outperforming PPO and heuristics in sum-rate simulations.
-
Addressing Overthinking in Large Vision-Language Models via Gated Perception-Reasoning Optimization
GPRO trains a meta-controller on 790k failure-labeled samples to dynamically select fast, perception, or reasoning paths in LVLMs, yielding higher accuracy and shorter responses than prior slow-thinking methods.
-
Mind the Generative Details: Direct Localized Detail Preference Optimization for Video Diffusion Models
LocalDPO aligns text-to-video diffusion models with human preferences at the spatio-temporal region level by automatically generating localized preference pairs from corrupted real videos and applying a region-aware DPO loss.
-
CORE: Concept-Oriented Reinforcement for Bridging the Definition-Application Gap in Mathematical Reasoning
CORE is a concept-oriented RL method that synthesizes quizzes, injects concept snippets into rollouts, and reinforces conceptual trajectories to close the gap between restating definitions and applying them in math problems.
-
Simultaneous Secrecy and Covert Communications (SSACC) in Mobility-Aware RIS-Aided Networks
A new SSACC scheme in mobility-aware RIS networks balances secrecy capacity and detection error probability via a GDM-DRL power allocation algorithm.
-
HetRL: Efficient Reinforcement Learning for LLMs in Heterogeneous Environments
HetRL delivers up to 9.17x higher throughput for LLM RL training on heterogeneous GPUs by using hybrid and ILP-based schedulers to solve a joint optimization problem over computation and data dependencies.
-
PIAST: Rapid Prompting with In-context Augmentation for Scarce Training data
PIAST iteratively optimizes few-shot examples in prompts via Monte Carlo Shapley value estimation, outperforming prior automatic prompting methods and setting new SOTA on classification, simplification, and GSM8K with modest compute.
-
Explicit Control Barrier Function-based Safety Filters and their Resource-Aware Computation
The paper gives explicit closed-form controllers for control barrier function safety filters via state-space partitioning and a switching implementation that recomputes only on region changes.
-
Asking like Socrates: Socrates helps VLMs understand remote sensing images
RS-EoT uses a SocraticAgent self-play system and two-stage RL to train VLMs for genuine iterative reasoning and visual inspection on remote sensing VQA and grounding tasks, achieving SOTA results.
-
VIDEOP2R: Video Understanding from Perception to Reasoning
VideoP2R separates perception and reasoning in a process-aware RFT pipeline with a new CoT dataset and PA-GRPO rewards, reaching SOTA on six of seven video benchmarks.
-
MURPHY: Feedback-Aware GRPO with Retrospective Credit Assignment for Multi-Turn Code Generation
MURPHY improves code generation pass rates by up to 6% through retrospective credit assignment on multi-turn feedback trees using max or mean reward propagation.
-
Q-RAG: Long Context Multi-step Retrieval via Value-based Embedder Training
Q-RAG trains embedders via RL for multi-step retrieval and reports state-of-the-art results on BabiLong and RULER benchmarks for contexts up to 10M tokens.
-
MemSearcher: Training LLMs to Reason, Search and Manage Memory via End-to-End Reinforcement Learning
MemSearcher trains LLMs to manage compact memory in multi-turn searches via multi-context GRPO for end-to-end RL, outperforming ReAct-style baselines with stable token counts.
-
On the Sample Complexity of Differentially Private Policy Optimization
Differential privacy in policy optimization adds sample complexity costs that often appear as lower-order terms rather than dominating the bounds.
-
CodeRL+: Improving Code Generation via Reinforcement with Execution Semantics Alignment
CodeRL+ integrates variable-level execution trajectory inference into RLVR training to align textual code representations with execution semantics, delivering 4.6% relative pass@1 gains and generalization to code-reasoning and test-output tasks.
-
The Art of Scaling Reinforcement Learning Compute for LLMs
A 400k+ GPU-hour study shows RL scaling in LLMs follows predictable sigmoidal trajectories, with most design choices affecting efficiency rather than the performance asymptote, enabling accurate large-scale predictions via the ScaleRL recipe.
-
HiPRAG: Hierarchical Process Rewards for Efficient Agentic Retrieval Augmented Generation
HiPRAG adds hierarchical process rewards to RL training for agentic RAG, reducing over-search to 2.3% and achieving 65.4-67.2% accuracy on seven QA benchmarks across 3B and 7B models.
-
Beyond Noisy-TVs: Noise-Robust Exploration Via Learning Progress Monitoring
LPM uses a dual-network design to compute intrinsic rewards from the change in prediction error across iterations, providing a noise-robust signal that is theoretically linked to information gain.
-
ZeroSiam: An Efficient Asymmetry for Test-Time Entropy Optimization without Collapse
ZeroSiam is an asymmetric architecture using a learnable predictor and stop-gradient that prevents collapse in test-time entropy minimization while also regularizing biased signals for improved performance.
-
DiffusionNFT: Online Diffusion Reinforcement with Forward Process
DiffusionNFT performs online RL for diffusion models on the forward process via flow matching and positive-negative contrasts, delivering up to 25x efficiency gains and rapid benchmark improvements over prior reverse-process methods.
-
Safe Reinforcement Learning using Action Projection: Safeguard the Policy or the Environment?
Action aliasing from safety projections harms policy-gradient estimates more severely when the projection is inside the policy than when it is outside, but a penalty term restores competitiveness.
-
Token Buncher: Shielding LLMs from Harmful Reinforcement Learning Fine-Tuning
TokenBuncher constrains response entropy via entropy-as-reward RL and a Token Noiser to stop harmful RL fine-tuning while keeping benign performance intact.
-
EyeMulator: Improving Code Language Models by Mimicking Human Visual Attention
EyeMulator augments CodeLLM fine-tuning loss with token weights derived from human eye-tracking scan paths, producing large gains on code translation and summarization across StarCoder, Llama-3.2 and DeepSeek-Coder.
-
BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion
BeyondMimic combines compact motion tracking with a unified guided latent diffusion model to master diverse agile behaviors from human demos and solve unseen downstream tasks via test-time classifier guidance.
-
MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE
MixGRPO speeds up GRPO for flow-based image generators by restricting SDE sampling and optimization to a sliding window while using ODE elsewhere, cutting training time by up to 71% with better alignment performance.
-
One Step is Enough: Multi-Agent Reinforcement Learning based on One-Step Policy Optimization for Order Dispatch on Ride-Sharing Platforms
OSPO trains optimal order dispatch policies for homogeneous AV fleets using only one-step group rewards, outperforming GRPO on a real ride-hailing dataset.
-
EXPO: Stable Reinforcement Learning with Expressive Policies
EXPO stabilizes online RL for expressive policies by training a base policy with imitation and using a lightweight Gaussian edit policy to select higher-value actions on the fly for sampling and TD backups.
-
High-Resolution Visual Reasoning via Multi-Turn Grounding-Based Reinforcement Learning
MGPO elicits grounding in LMMs via multi-turn RL with binary rewards, yielding 5.4% and 5.2% gains on MME-Realworld and V* Bench and surpassing GPT-4o on the latter after training on 21K samples.
-
Preemptive Solving of Future Problems: Multitask Preplay in Humans and Machines
Multitask Preplay replays experience from pursued tasks as starting points for counterfactual simulation of unpursued tasks to learn predictive representations that support fast generalization in humans and machines.
-
Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling
Prefix-RFT blends SFT and RFT via prefix sampling from demonstrations to outperform standalone SFT, RFT, and mixed-policy baselines on math reasoning problems.
-
MMSearch-R1: Incentivizing LMMs to Search
MMSearch-R1 uses reinforcement learning to train multimodal models for on-demand multi-turn internet search with image and text tools, outperforming same-size RAG baselines and matching larger ones while cutting search calls by over 30%.