Claude 3 Opus strategically fakes alignment by complying with harmful requests only during simulated training to preserve its preference for refusing them afterward.
super hub Canonical reference
Proximal Policy Optimization Algorithms
Canonical reference. 78% of citing Pith papers cite this work as background.
692
Pith papers citing it
Background
78% 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 7
method 2
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
co-cited works
representative citing papers
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.
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.
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.
SubMAPG uses a new Partition Multilinear Extension to derive unbiased policy gradients from submodular difference rewards, delivering 1/2-approximation and sublinear dynamic regret for online distributed task allocation in open multi-agent systems.
F-GRPO factorizes group-relative policy optimization into generation and ranking phases within one autoregressive sequence, using order-invariant coverage and position-aware utility rewards to improve top-ranked performance on recommendation and multi-hop QA tasks.
An adaptive smooth Tchebycheff controller for multi-objective RL lets agents reach non-convex Pareto regions in robotic tasks while avoiding the instability of static non-linear scalarizations.
LLM simulators exhibit near-zero selective response to targeted misconception feedback and behave sycophantically, but SFT and SFS-aligned RL improve this property.
MOPD improves on-policy distillation for LLMs by using peer successes for positive patterns and failures for negative examples to create more informative teacher signals.
QAP-Router models qubit routing as dynamic QAP and applies RL with a solution-aware Transformer to cut CNOT counts by 12-30% versus industry compilers on real circuit benchmarks.
Miss-MDPs extend POMDPs with missing-data theory to learn observation missingness patterns and compute near-optimal policies with high-probability guarantees.
dGRPO merges outcome-based policy optimization with dense teacher guidance from on-policy distillation, yielding more stable long-context reasoning on the new LongBlocks synthetic dataset.
Multistability is necessary for temporal horizon generalization in POMDPs, sufficient in simple tasks along with transient dynamics in complex ones, while monostable parallelizable RNNs like SSMs and gated linear RNNs fail by construction.
ActGuide-RL uses human action data as plan-style guidance in mixed-policy RL to overcome exploration barriers in LLM agents, matching SFT+RL performance on search benchmarks without cold-start training.
DGAO uses reinforcement learning to optimize LLMs for both accuracy and order stability by balancing intra-group accuracy advantages and inter-group stability advantages.
StepCodeReasoner aligns code reasoning with verifiable stepwise execution traces via print anchors and bi-level GRPO reinforcement learning, reaching SOTA results on CRUXEval (91.1%) and LiveCodeBench (86.5%) for a 7B model.
Delightful Policy Gradient removes exponential corner trapping in softmax policy optimization for bandits and tabular MDPs, achieving logarithmic escape times and global O(1/t) convergence.
Anchor-guided variance-aware reward modeling uses two response-level anchors to resolve non-identifiability in Gaussian models of pluralistic preferences, yielding provable identification, a joint training objective, and improved RLHF performance.
citing papers explorer
-
ReLibra: Routing-Replay-Guided Load Balancing for MoE Training in Reinforcement Learning
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 Models)
Weak-to-strong generalization is nearly inevitable in linear logistic regression for most student-teacher pairs without any model capacity mismatch.
-
Structural Equivalence and Learning Dynamics in Delayed MARL
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.
-
F-GRPO: Factorized Group-Relative Policy Optimization for Unified Candidate Generation and Ranking
F-GRPO factorizes group-relative policy optimization into generation and ranking phases within one autoregressive sequence, using order-invariant coverage and position-aware utility rewards to improve top-ranked performance on recommendation and multi-hop QA tasks.
-
Multi-Rollout On-Policy Distillation via Peer Successes and Failures
MOPD improves on-policy distillation for LLMs by using peer successes for positive patterns and failures for negative examples to create more informative teacher signals.
-
On the Importance of Multistability for Horizon Generalization in Reinforcement Learning
Multistability is necessary for temporal horizon generalization in POMDPs, sufficient in simple tasks along with transient dynamics in complex ones, while monostable parallelizable RNNs like SSMs and gated linear RNNs fail by construction.
-
Towards Order Fairness: Mitigating LLMs Order Sensitivity through Dual Group Advantage Optimization
DGAO uses reinforcement learning to optimize LLMs for both accuracy and order stability by balancing intra-group accuracy advantages and inter-group stability advantages.
-
Delightful Gradients Accelerate Corner Escape
Delightful Policy Gradient removes exponential corner trapping in softmax policy optimization for bandits and tabular MDPs, achieving logarithmic escape times and global O(1/t) convergence.
-
Block-R1: Rethinking the Role of Block Size in Multi-domain Reinforcement Learning for Diffusion Large Language Models
Introduces Block-R1 benchmark, Block-R1-41K dataset, and a conflict score to handle domain-specific optimal block sizes in RL post-training of diffusion LLMs.
-
gym-invmgmt: An Open Benchmarking Framework for Inventory Management Methods
gym-invmgmt is a new benchmarking framework that evaluates inventory policies across optimization and learning methods, finding stochastic programming strongest among non-oracle approaches and PPO-Transformer best among learned ones in tested scenarios.
-
Learning When to Act: Communication-Efficient Reinforcement Learning via Run-Time Assurance
Learned policies with runtime Lyapunov shields achieve substantially higher communication intervals than baselines while maintaining stability on inverted pendulum, cart-pole, and quadrotor systems.
-
Newton's Lantern: A Reinforcement Learning Framework for Finetuning AC Power Flow Warm Start Models
Newton's Lantern is an RL finetuning pipeline that uses iteration count as reward to produce warm starts for AC power flow, outperforming supervised methods by converging on all tested snapshots with lowest mean iterations on IEEE and GOC benchmarks.
-
Natural Policy Gradient as Doubly Smoothed Policy Iteration: A Bellman-Operator Framework
Natural policy gradient is a special case of doubly smoothed policy iteration that achieves distribution-free global geometric convergence to an epsilon-optimal policy in O((1-gamma)^{-1} log((1-gamma)^{-1} epsilon^{-1})) iterations.
-
Controllability in preference-conditioned multi-objective reinforcement learning
Standard MORL metrics do not measure whether preference inputs reliably control agent behavior, so a new controllability metric is introduced to restore the link between user intent and agent output.
-
Trust Region Inverse Reinforcement Learning: Explicit Dual Ascent using Local Policy Updates
TRIRL enables explicit dual-ascent IRL via trust-region local policy updates that guarantee monotonic improvement without full RL solves per iteration, outperforming prior imitation methods by 2.4x aggregate IQM and recovering generalizable rewards.
-
Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
-
Revisiting Mixture Policies in Entropy-Regularized Actor-Critic
A new marginalized reparameterization estimator allows low-variance training of mixture policies in entropy-regularized actor-critic algorithms, matching or exceeding Gaussian policy performance in several continuous control benchmarks.
-
Non-Parametric Rehearsal Learning via Conditional Mean Embeddings
A non-parametric rehearsal learning framework using conditional mean embeddings and a Probit surrogate for avoiding undesired outcomes, with consistency guarantees.
-
BubbleSpec: Turning Long-Tail Bubbles into Speculative Rollout Drafts for Synchronous Reinforcement Learning
BubbleSpec exploits long-tail bubbles in synchronous RL by using faster ranks' idle time to pre-generate rollout drafts for speculative decoding, reducing steps by 50% and raising throughput up to 1.8x while preserving exact synchrony.
-
TMPO: Trajectory Matching Policy Optimization for Diverse and Efficient Diffusion Alignment
TMPO uses Softmax Trajectory Balance to match policy probabilities over multiple trajectories to a Boltzmann reward distribution, improving diversity by 9.1% in diffusion alignment tasks.
-
The Cancellation Hypothesis in Critic-Free RL: From Outcome Rewards to Token Credits
The cancellation hypothesis shows how rollout-level rewards produce token-level credit assignment in critic-free RL through cancellation of opposing signals on shared tokens, with empirical support and batching interventions that enhance performance.
-
KL for a KL: On-Policy Distillation with Control Variate Baseline
vOPD stabilizes on-policy distillation gradients by subtracting a closed-form per-token negative reverse KL baseline as a detached control variate, preserving unbiasedness while lowering variance and matching expensive full-vocabulary methods.
-
Approximation-Free Differentiable Oblique Decision Trees
DTSemNet gives an exact, invertible neural-network encoding of hard oblique decision trees that supports direct gradient training for both classification and regression without probabilistic softening or quantized estimators.
-
Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States
POISE trains a lightweight probe on the actor's internal states to predict expected rewards for RLVR, matching DAPO performance on math benchmarks with lower compute by avoiding extra rollouts or critic models.
-
Convex Optimization with Nested Evolving Feasible Sets
For convex losses in nested evolving feasible sets, a lazy algorithm balances O(T^{1-β}) regret with O(T^β) movement for any β; for strongly convex or sharp losses, Frugal achieves zero regret with O(log T) movement, shown optimal by matching lower bound.
-
Rethinking Importance Sampling in LLM Policy Optimization: A Cumulative Token Perspective
The cumulative token IS ratio gives unbiased prefix correction and lower variance than full-sequence ratios for token-level gradients in LLM policy optimization, enabling CTPO to outperform GRPO and GSPO baselines on mathematical reasoning tasks.
-
Theoretical Limits of Language Model Alignment
The maximum reward gain under KL-regularized LM alignment is a Jeffreys divergence term, estimable as covariance from base samples, with best-of-N approaching the theoretical limit.
-
Rollback-Free Stable Brick Structures Generation
Reinforcement learning internalizes physical stability rules for brick structures, enabling the first rollback-free generation with orders-of-magnitude faster inference.
-
PairAlign: A Framework for Sequence Tokenization via Self-Alignment with Applications to Audio Tokenization
PairAlign learns compact audio token sequences via self-alignment of paired content views using an autoregressive decoder, achieving strong cross-view consistency and edit-distance preservation while reducing token count by 55% on TIMIT.
-
Operator-Guided Invariance Learning for Continuous Reinforcement Learning
VPSD-RL discovers exact and approximate value-preserving Lie-group operators in continuous RL to stabilize learning via transition augmentation and consistency regularization.
-
Does Synthetic Data Help? Empirical Evidence from Deep Learning Time Series Forecasters
Synthetic data augmentation helps channel-mixing time series models but degrades channel-independent ones, with reliable gains only from seasonal-trend generators and gradual schedules in low-resource settings.
-
Enabling Unsupervised Training of Deep EEG Denoisers With Intelligent Partitioning
iPSD enables self-supervised training of deep EEG denoisers by learning to partition noisy segments into independent noisy realizations of the same neural activity, achieving state-of-the-art performance at very low SNR without clean references.
-
Selective Rollout: Mid-Trajectory Termination for Multi-Sample Agent RL
A one-parameter early-termination gate based on mean pairwise prefix edit distance reduces wall-clock time by 10.7% and raises held-out success by 2.5 pp in GRPO on ALFWorld by cutting zero-advantage batch dilution.
-
Non-Myopic Active Feature Acquisition via Pathwise Policy Gradients
NM-PPG optimizes non-myopic acquisition policies for costly features by enabling pathwise gradients via continuous relaxation and straight-through rollouts in POMDPs, outperforming SOTA baselines.
-
Approximate Next Policy Sampling: Replacing Conservative Target Policy Updates in Deep RL
Approximate Next Policy Sampling approximates the next policy's state distribution during training to enable larger safe policy updates in deep RL, demonstrated by SV-PPO matching or exceeding standard PPO on Atari and continuous control tasks.
-
FLUID: Continuous-Time Hyperconnected Sparse Transformer for Sink-Free Learning
FLUID is a continuous-time transformer using Liquid Attention Networks to model attention as stable ODE solutions that interpolate between discrete SDPA and CT-RNNs, with an explicit sink gate and liquid hyper-connections for better information flow.
-
Reference-Sampled Boltzmann Projection for KL-Regularized RLVR: Target-Matched Weighted SFT, Finite One-Shot Gaps, and Policy Mirror Descent
Reference-sampled weighted SFT with prompt-normalized Boltzmann weights induces the same policy as fixed-reference KL-regularized RLVR, with BOLT as the estimator and a finite one-shot error decomposition separating coverage, variance, and other terms.
-
PACE: Parameter Change for Unsupervised Environment Design
PACE uses the squared L2 norm of policy parameter changes from a first-order approximation as an efficient proxy for environment value in UED, outperforming baselines with higher IQM and lower optimality gap on MiniGrid and Craftax OOD tests.
-
Forager: a lightweight testbed for continual learning with partial observability in RL
Forager is a lightweight partially-observable continual RL environment that exposes loss of plasticity in current agents and highlights the value of state construction for ongoing learning.
-
Deep Variational Inference Symbolic Regression
DVISR performs variational inference over symbolic expression trees and constants by training a neural network with the ELBO as reward, recovering true posteriors in simple test cases.
-
InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees
InvEvolve evolves white-box inventory policies from LLMs with statistical safety guarantees and outperforms classical and deep learning methods on synthetic and real retail data.
-
Uniform-Correct Policy Optimization: Breaking RLVR's Indifference to Diversity
UCPO modifies GRPO with a uniformity penalty over correct solutions to prevent diversity collapse in RLVR, yielding up to 10% higher Pass@64 on AIME24 and 45% more equation-level diversity.
-
BoostLoRA: Growing Effective Rank by Boosting Adapters
BoostLoRA grows effective adapter rank linearly via iterative boosting on hard examples with orthogonal low-rank updates, outperforming both single-shot ultra-low-rank adapters and full fine-tuning on math and code tasks with zero added inference overhead.
-
GradMAP: Gradient-Based Multi-Agent Proximal Learning for Grid-Edge Flexibility
GradMAP enables fast offline training of fully decentralized neural policies for grid-edge flexibility by embedding a differentiable three-phase AC power-flow model and applying proximal surrogates in action space.
-
R2IF: Aligning Reasoning with Decisions via Composite Rewards for Interpretable LLM Function Calling
R2IF improves LLM function-calling accuracy by up to 34.62% on BFCL using a composite reward system with CER and SMV components optimized via GRPO, while increasing interpretability through positive CoT effectiveness.
-
EVPO: Explained Variance Policy Optimization for Adaptive Critic Utilization in LLM Post-Training
EVPO adaptively switches between critic-based and batch-mean advantage estimation using batch-level explained variance to provably achieve no greater variance than the better of PPO or GRPO at every step.
-
Guiding Distribution Matching Distillation with Gradient-Based Reinforcement Learning
GDMD replaces raw-sample rewards with distillation-gradient rewards in RL-guided diffusion distillation, yielding 4-step models that surpass their multi-step teachers on GenEval and human preference metrics.
-
Self-Improving Tabular Language Models via Iterative Group Alignment
TabGRAA enables self-improving tabular language models through iterative group-relative advantage alignment using modular automated quality signals like distinguishability classifiers.
-
Bounded Ratio Reinforcement Learning
BRRL derives an analytic optimal policy for regularized constrained RL that guarantees monotonic improvement and yields the BPO algorithm that matches or exceeds PPO.
-
Rethinking the Comparison Unit in Sequence-Level Reinforcement Learning: An Equal-Length Paired Training Framework from Loss Correction to Sample Construction
EqLen is a sample-construction framework that builds equal-length paired segments via dual-track generation and masking for stable group-relative RL in sequences, reframing the length problem as a comparison-unit issue rather than loss scaling.