ICLR 2026PastLarge language models
The 1st Workshop on Scaling Post-training for LLMs
SPOT
- Submission deadline
- Feb 7, 2026, 11:59 UTCOpenReview-synced 2026-02-07 11:59 UTC (as of 2026-06-23) — extensions on OpenReview are applied automatically; verify on the website.
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (64)
Fetched from OpenReview (v2) on 2026-06-10.
A-3PO: Accelerating Asynchronous LLM Training with Staleness-aware Proximal Policy Approximation
Actor-Curator: Scalable Adaptive Curriculum Learning for LLM Post-Training
Beyond Log Likelihood: Probability-Based Objectives for Supervised Fine-Tuning across the Model Capability Continuum
Beyond Scalar Critics: A Distributional Perspective on Reinforcement Learning with Verifiable Rewards for LLMs
BugPilot: Complex Bug Generation for Efficient Learning of SWE Skills
Challenges in Inference-Time Scaling with Uncertainty-Aware Tree Search
CLEANER: Self-Purified Trajectories Boost Agentic Reinforcement Learning
Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
Compute-Efficient GRPO Training
Corecraft: Training Generalizable Agents on High-Fidelity RL Environments
Countdown-Code: A Testbed for Studying The Emergence and Generalization of Reward Hacking
Counterfactual Credit Assignment for Policy Optimization
Coverage Improvement and Fast Convergence of On-policy Preference Learning
CrispEdit: Low-Curvature Projections for Scalable Non-Destructive LLM Editing
DELTA4: Sparse Matrix-Vector Multiplication for Low Sparsity
DGPO: Decoupled Gradient Policy Optimization for RLVR in LLMs
DIRICHLET-PRIOR SHAPING: GUIDING EXPERT SPECIALIZATION IN UPCYCLED MOES
Efficient and Stable Scaling of Reinforcement Learning for LLMs via Dynamic Allocation and Gradient Modulation
Efficient RL Training for LLMs with Experience Replay
Entropy-Aware On-Policy Distillation of Language Models
Escaping the Mode: Multi Answer Reinforcement Learning in LMs
Execution-Grounded Credit Assignment for GRPO in Code Generation
Expanding the Capabilities of Reinforcement Learning via Text Feedback
F-GRPO: Don't Let Your Policy Learn the Obvious and Forget the Rare
Federated Agent Reinforcement Learning
From Atomic to Composite: Reinforcement Learning Enables Generalization in Complementary Reasoning
GEOMA: Geometric and Econometric Objectives for Multi-Reward Alignment
Hierarchical Agenda Reasoning for Strategic Multi-Turn Dialogue Agents
Is the Importance Ratio Necessary for Stable Reinforcement Learning in LLMs?
IsoCompute Playbook: Optimally Scaling Sampling Compute for LLM RL
Jointly Reinforcing Diversity and Quality in Language Model Generations
Learning Discriminative Process Reward Models without Step Labels
Learning Useful Supervision for Reinforcement Learning in Reasoning Models
Making Complex Reasoning Student-Friendly: A Hybrid LLM-to-SLM Distillation Framework
Maximum Likelihood Reinforcement Learning
Mix Early, Forget Less: Data Mixing During Pretraining Builds Resistance to Forgetting
Near-Optimal Regret for KL-Regularized Multi-Armed Bandits
NyoomFloat12: Lossless 12-bit Weight Compression for Post-Training Inference
On quantizing the state of the Muon optimizer
Prescriptive Scaling Reveals the Evolution of Language Model Capabilities
Privileged Information Distillation for Language Models
QFT: Quantized Full-parameter Tuning of LLMs with Affordable Resources
Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneck
Reasoning Cache: Learning to Extrapolate to Long Lengths via Short-Length RL
Recontextualization Mitigates Specification Gaming without Modifying the Specification
Reinforcement Learning via Self-Distillation
Reuse your FLOPs: Scaling RL on Hard Problems by Conditioning on Very Off-Policy Prefixes
RewardFlow: Topology-Aware Reward Propagation on State Graphs for Agentic RL with Large Language Models
RL Excursions during Pre-training: How early is too early for On-policy Learning?
RL-VLA$^3$: Reinforcement Learning VLA Accelerating via Full Asynchronism
Scaling Reward Modeling without Human Supervision
Scaling Search-Augmented LLM Reasoning via Adaptive Information Control
Semantic Tube Prediction: Beating LLM Data Efficiency with JEPA
Shape of Thought: When Distribution Matters More than Correctness in Reasoning Tasks
Sparse Attention for Efficient LLM Reinforcement Learning
TestSmith: Reinforcement Learning for Unit Test Generation with Synthetic Perturbations
TOUCAN: Synthesizing 1.5M Tool-Agentic Data from Real-World MCP Environments
Towards Understanding the Benefits of Online Imitation Learning
Training-Free Dynamic Upcycling of Expert Language Models
V1: Unifying Generation and Self-Verification for Parallel Reasoners
VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use
Weight Decay Improves Language Model Plasticity
Weight Space Detection of Backdoors in LoRA Adapters
When Tokens Decay and Turns Amplify: A Dual-Granularity Framework for Multi-Turn Preference Optimization