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%).
1591
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.
A reward-free representation learning pipeline for offline PbRL achieves better preference efficiency than standard two-stage baselines by connecting RFRL concepts to preference data.
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.
citing papers explorer
-
Mechanistic Analysis of Alignment Algorithms in Language Models
Mechanistic analysis of six preference optimization methods reveals distinct geometric shifts in model representations, with KTO/GRPO enhancing separability while DPO/ORPO degrade it.
-
Structured Recurrent Mixers for Massively Parallelized Sequence Generation
Structured Recurrent Mixers provide a dual parallel-recurrent representation for sequence models, claiming superior training efficiency, information capacity, and inference throughput over linear complexity alternatives.
-
Iterative Critique-and-Routing Controller for Multi-Agent Systems with Heterogeneous LLMs
A critique-and-routing controller cast as a finite-horizon MDP with policy-gradient optimization outperforms one-shot routing baselines on reasoning benchmarks while using the strongest agent for under 25% of calls.
-
$\xi$-DPO: Direct Preference Optimization via Ratio Reward Margin
ξ-DPO rewrites the preference objective as minimizing distance to optimal margins and defines reward as a chosen-to-rejected ratio, yielding a bounded, interpretable margin ξ set directly from the initial reward-gap distribution.
-
Rotation-Preserving Supervised Fine-Tuning
RPSFT improves the in-domain versus out-of-domain performance trade-off during LLM supervised fine-tuning by penalizing rotations in pretrained singular subspaces as a proxy for loss-sensitive directions.
-
AIPO: Learning to Reason from Active Interaction
AIPO adds active multi-agent consultation (Verify, Knowledge, Reasoning agents) plus custom importance sampling to RLVR training so LLMs expand their reasoning boundary and then operate without the agents.
-
Rubric-Grounded RL: Structured Judge Rewards for Generalizable Reasoning
Rubric-grounded RL with LLM judges on document-derived criteria raises Llama-3.1-8B normalized reward to 71.7% on held-out rubrics and improves performance on GSM8K, MATH, and GPQA benchmarks.
-
Beyond Pairs: Your Language Model is Secretly Optimizing a Preference Graph
GraphDPO generalizes pairwise DPO to a graph-structured Plackett-Luce objective over DAGs induced by rollout rankings, enforcing transitivity with linear complexity and recovering DPO as a special case.
-
Actor-Critic Algorithm for Dynamic Expectile and CVaR
A model-free off-policy actor-critic algorithm is constructed for dynamic expectile and CVaR using a surrogate policy gradient without transition perturbation and elicitability-based value learning, with empirical outperformance in risk-averse domains.
-
Reflective Prompted Policy Optimization: Trajectory-Grounded Revision and Salience Bias
Reflective Prompted Policy Optimization uses a Critic-LLM to inspect full trajectories and propose grounded revisions, yielding higher mean best rewards, faster near-optimal performance, and greater stability than scalar-reward baselines across ten environments.
-
Drifting Field Policy: A One-Step Generative Policy via Wasserstein Gradient Flow
DFP is a one-step generative policy using Wasserstein gradient flow on a drifting model backbone, with a top-K behavior cloning surrogate, that reaches SOTA on Robomimic and OGBench manipulation tasks.
-
SOD: Step-wise On-policy Distillation for Small Language Model Agents
SOD reweights on-policy distillation strength step-by-step using divergence to stabilize tool use in small language model agents, yielding up to 20.86% gains and 26.13% on AIME 2025 for a 0.6B model.
-
Gradient Starvation in Binary-Reward GRPO: Why Group-Mean Centering Fails and Why the Simplest Fix Works
Group-mean centering in binary-reward GRPO produces gradient starvation; the fixed sign advantage A=2r-1 raises GSM8K accuracy from 28.4% to 73.8% at group size 4.
-
WeatherSyn: An Instruction Tuning MLLM For Weather Forecasting Report Generation
WeatherSyn is the first instruction-tuned MLLM for weather forecasting report generation, outperforming closed-source models on a new dataset of 31 US cities across 8 weather aspects.
-
HTPO: Towards Exploration-Exploitation Balanced Policy Optimization via Hierarchical Token-level Objective Control
HTPO introduces hierarchical token-level objective control in RLVR to balance exploration and exploitation by grouping tokens according to difficulty, correctness, and entropy, yielding up to 8.6% gains on AIME benchmarks over DAPO.
-
RELO: Reinforcement Learning to Localize for Visual Object Tracking
RELO formulates visual object tracking localization as a Markov decision process solved by reinforcement learning with combined IoU and AUC rewards, augmented by layer-aligned temporal token propagation, and reports 57.5% AUC on LaSOText without template updates.
-
Experience Sharing in Mutual Reinforcement Learning for Heterogeneous Language Models
Mutual Reinforcement Learning allows heterogeneous LLMs to exchange experience through mechanisms like Peer Rollout Pooling, Cross-Policy GRPO Advantage Sharing, and Success-Gated Transfer, with outcome-level sharing identified as favorable on the stability-support trade-off.
-
Can You Break RLVER? Probing Adversarial Robustness of RL-Trained Empathetic Agents
RLVER agents improve emotional responsiveness under adversarial user behaviors but exhibit no measurable gains in tracking emotional states compared to untuned base models.
-
Dr. Post-Training: A Data Regularization Perspective on LLM Post-Training
Dr. Post-Training reframes general data as a data-induced regularizer for LLM post-training updates, yielding a family of methods that outperform data-selection baselines on SFT, RLHF, and RLVR tasks.
-
Response Time Enhances Alignment with Heterogeneous Preferences
Response times modeled as drift-diffusion processes enable consistent estimation of population-average preferences from heterogeneous anonymous binary choices.
-
Bridging Textual Profiles and Latent User Embeddings for Personalization
BLUE aligns LLM-generated textual user profiles with embedding-based recommendation objectives via reinforcement learning and next-item text supervision, yielding better zero-shot performance and cross-domain transfer than baselines.
-
Sequential Design of Genetic Circuits Under Uncertainty With Reinforcement Learning
An amortized reinforcement learning method enables immediate, observation-driven sequential optimization of genetic circuits while accounting for both intrinsic stochasticity and cross-laboratory variability without repeated inference steps.
-
Preference Instability in Reward Models: Detection and Mitigation via Sparse Autoencoders
Sparse autoencoders isolate unstable features in reward model representations and enable two mitigation techniques that reduce preference errors on perturbed inputs without retraining.
-
Process Matters more than Output for Distinguishing Humans from Machines
A new battery of 30 cognitive tasks demonstrates that process-level behavioral features distinguish humans from frontier AI agents better than performance metrics (mean AUC 0.88), with process-specific fine-tuning improving mimicry but limited cross-task transfer.
-
Gradient Extrapolation-Based Policy Optimization
GXPO approximates longer local lookahead in GRPO training via gradient extrapolation from two optimizer steps using three backward passes total, improving pass@1 accuracy by 1.65-5.00 points over GRPO and delivering up to 4x step speedup.
-
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.
-
Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
Listwise Policy Optimization explicitly performs target-projection on the LLM response simplex, unifying and improving group-based RLVR methods with monotonic improvement and flexible divergences.
-
Optimal Transport for LLM Reward Modeling from Noisy Preference
SelectiveRM applies optimal transport with a joint consistency discrepancy and partial mass relaxation to produce reward models that optimize a tighter upper bound on clean risk while autonomously dropping noisy preference samples.
-
BehaviorGuard: Online Backdoor Defense for Deep Reinforcement Learning
BehaviorGuard detects backdoor behaviors in DRL policies via behavioral drift in action distributions and suppresses suspicious actions at runtime, claimed as the first online defense for both single- and multi-agent settings.
-
Beyond Uniform Credit Assignment: Selective Eligibility Traces for RLVR
S-trace adds sparse eligibility traces to RLVR that mask low-entropy tokens, outperforming GRPO by 0.49-3.16% pass@16 on Qwen3 models while improving sample and token efficiency.
-
DexSynRefine: Synthesizing and Refining Human-Object Interaction Motion for Physically Feasible Dexterous Robot Actions
DexSynRefine couples HOI motion manifold flow primitives with task-space residual RL and proprioceptive adaptation to convert human-object interaction data into executable dexterous robot motions, reporting 50-70 point real-world success rate gains over kinematic retargeting on five tasks.
-
Think, then Score: Decoupled Reasoning and Scoring for Video Reward Modeling
DeScore decouples CoT reasoning from reward scoring in video reward models using a two-stage training process to improve generalization and avoid optimization bottlenecks of coupled generative RMs.
-
OpenG2G: A Simulation Platform for AI Datacenter-Grid Runtime Coordination
OpenG2G is a new extensible simulation platform that lets users implement and compare classic, optimization, and learning-based controllers for AI datacenter power flexibility coordinated with the grid.
-
Rollout Pass-Rate Control: Steering Binary-Reward RL Toward Its Most Informative Regime
Prefix Sampling replays self-generated trajectory prefixes to control rollout pass rates near 50% in binary-reward RL, delivering wall-clock speedups and modest performance gains on SWE-bench Verified and AIME tasks.
-
LineRides: Line-Guided Reinforcement Learning for Bicycle Robot Stunts
LineRides enables commandable bicycle robot stunts via line-guided RL that uses spatial guidelines, a tracking margin for feasibility, distance-based progress, and sparse key-orientations.
-
Reward-Decomposed Reinforcement Learning for Immersive Video Role-Playing
EBM-RL applies a GRPO-based RL method with decomposed rewards for scene alignment, perceptual utility, faithfulness, and format to improve video-grounded role-playing dialogue over text-only baselines.
-
Threshold-Guided Optimization for Visual Generative Models
A threshold-guided alignment method lets visual generative models be optimized directly from scalar human ratings instead of requiring paired preference data.
-
Counter-Dyna: Data-Efficient RL-Based HVAC Control using Counterfactual Building Models
Counter-Dyna reduces RL training data for HVAC control to five weeks by using counterfactual surrogate models that ignore uncontrollable variables like weather and prices.
-
RLearner-LLM: Balancing Logical Grounding and Fluency in Large Language Models via Hybrid Direct Preference Optimization
RLearner-LLM achieves up to 6x gains in NLI entailment over standard fine-tuning by using an automated hybrid DPO pipeline that balances logic and fluency across multiple model sizes and domains.
-
One Pool, Two Caches: Adaptive HBM Partitioning for Accelerating Generative Recommender Serving
HELM adaptively partitions HBM between EMB and KV caches via a three-layer PPO controller and EMB-KV-aware scheduling, reducing P99 latency by 24-38% while achieving 93.5-99.6% SLO satisfaction on production workloads.
-
Constraint-Enhanced Reinforcement Learning Based on Dynamic Decoupled Spherical Radial Squashing
DD-SRad is a new RL constraint technique that adapts per-actuator radii dynamically to achieve zero violations and unconstrained-level task performance on heterogeneous robotic joints.
-
Mitigating False Positives in Static Memory Safety Analysis of Rust Programs via Reinforcement Learning
Reinforcement learning on MIR features combined with cargo-fuzz validation reduces false positives in Rust static memory safety analysis, raising precision from 25.6% to 59.0% and accuracy to 65.2%.
-
SigLoMa: Learning Open-World Quadrupedal Loco-Manipulation from Ego-Centric Vision
SigLoMa enables dynamic loco-manipulation on quadrupeds from ego-centric 5 Hz vision alone by using Sigma Points for scalable exteroception, an ego-centric Kalman Filter for high-rate state estimation, and an active sampling curriculum, matching expert human teleoperation performance.
-
SOAR: Real-Time Joint Optimization of Order Allocation and Robot Scheduling in Robotic Mobile Fulfillment Systems
SOAR is a unified DRL method using soft allocations, event-driven MDP, and heterogeneous graph transformers that cuts global makespan by 7.5% and average order completion time by 15.4% at sub-100ms latency in RMFS.
-
CASCADE: Case-Based Continual Adaptation for Large Language Models During Deployment
CASCADE enables LLMs to continually adapt at deployment via case-based episodic memory and contextual bandits, improving macro-averaged success by 20.9% over zero-shot on 16 tasks spanning medicine, law, code, and robotics.
-
Stage Light is Sequence$^2$: Multi-Light Control via Imitation Learning
SeqLight maps music to multi-light HSV control via SkipBART for global color prediction followed by hybrid imitation learning in a goal-conditioned MDP to decompose colors across lights.
-
FINER-SQL: Boosting Small Language Models for Text-to-SQL
FINER-SQL boosts 3B-parameter small language models to 67.73% and 85% execution accuracy on BIRD and Spider benchmarks via dense memory and atomic rewards in group relative policy optimization, matching larger LLMs at lower latency.
-
Replacing Parameters with Preferences: Federated Alignment of Heterogeneous Vision-Language Models
MoR lets clients train local reward models on private preferences and uses a learned Mixture-of-Rewards with GRPO on the server to align a shared base VLM without exchanging parameters, architectures, or raw data.
-
Learning Reactive Dexterous Grasping via Hierarchical Task-Space RL Planning and Joint-Space QP Control
A multi-agent RL high-level planner outputs task-space velocities that a GPU-parallel QP low-level controller converts to joint velocities while enforcing limits and collisions, yielding robust sim-to-real dexterous grasping with zero-shot steerability.
-
DGPO: Distribution Guided Policy Optimization for Fine Grained Credit Assignment
DGPO is a critic-free RL framework that uses bounded Hellinger distance and entropy-gated advantage redistribution to enable fine-grained token-level credit assignment in long CoT generations for LLM alignment, reporting SOTA results on AIME benchmarks.