Rejection Sampling

Follow this topic Ask about this topic

Rejection sampling proposes candidates and accepts or discards them to obtain samples from a desired distribution, with proposal quality governing how much computation is wasted. Methods include learned proposals, threshold-free approximate samplers, and token- or segment-level rejection in language-model generation and post-training. Applications range from posterior approximation and matrix sampling to reducing hallucinations, improving reasoning data, and replanning after robot failures. The central tension is balancing efficiency and sample quality with guarantees about the resulting distribution, especially when acceptance criteria or training data introduce bias.

Latest papers

13 papers2 papers in the last four weeks, with none the four weeks before. 0.0% of all new papers.

Verdicts Without Annotated Evidence: Rejection Sampling or Label-Only Post-Training for Evidence Recovery?
Oct 3, 2026Nishanth Nayakanti, Prasang Gupta, Ashutosh Bilthare +1Language Model Post-Training Rejection Sampling
PricewaterhouseCoopers, U.S.
Uniform Race: Parameter-Free Approximate Rejection Sampling
Sep 28, 2026Seiyun Shin, Juhyeong Pang, Kwang-Sung JunRejection Sampling Sampling from Unnormalized Distributions
Graduate School of Artificial Intelligence Pohang University of Science and Technology Pohang, 37673, South Korea · Department of Computer Science University of Wisconsin–Madison Madison, WI 53706, USA · Graduate School of Artificial Intelligence Department of Computer Science and Engineering Pohang University of Science and Technology Pohang, 37673, South Korea
Fast Length-Squared Sampling for Positive-Semidefinite Matrices
Aug 12, 2026Rajarshi Bhattacharjee, Ethan N. Epperly, Cameron Musco +1Importance Sampling Rejection Sampling
University of Massachusetts Amherst · University of California Berkeley
Turning Off-Policy Tokens On-Policy: A Plug-in Approach for Improving LLM Alignment
Jul 6, 2026Yu Li, Xiuyu Li, Mingyang Yi +4LLM Alignment Importance Sampling
Renmin University of China
When Does Generating More Help? Disentangling Fixed-Source Synthesis from Source Expansion in Synthetic Data Scaling
Jul 2, 2026Xu Guo, Jian Tong, Zhihui Lu +1Synthetic Data Generation Language Model Scaling Laws
Shanghai AI Laboratory · Shanghai Innovation Institute · Fudan University
Diagnosing and Mitigating Context Rot in Long-horizon Search
Jun 29, 2026Shijie Xia, Yikun Wang, Zhen Huang +1Long-Context Modeling LLM Context Management
Shanghai Jiao Tong University · SII · GAIR +1
The Quality-Utility Paradox: Why High-Reward Data Impairs Small Model Mathematical Reasoning
Jun 15, 2026Haolong Qian, Xianliang Yang, Yinuo ma +6CoT Distillation Language Model Distillation
Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China · Microsoft Research Asia, Microsoft, Beijing, China · Faculty of Computer Science and Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China +1
Implicit Variational Rejection Sampling
Jun 12, 2026Jian Xu, Shigui Li, Wei Chen +6Bayesian Inference Variational Inference
RIKEN iTHEMS · RIKEN AIP · South China University of Technology +2
Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling
Jun 10, 2026Yucheng Li, Huiqiang Jiang, Yang Xu +14Multi-Token Prediction Speculative Decoding
Qwen Team, Alibaba Inc.
Building Reliable Long-Form Generation via Hallucination Rejection Sampling
Jun 2, 2026Lin Li, Georgia Channing, Suhaas M Bhat +2Long-Form Document Generation LLM Hallucination Mitigation
Department of Computer Science, University of Oxford, Oxford, UK · Department of Engineering, University of Oxford, Oxford, UK · Oxford Digital Health Labs, Nuffield Department of Women’s and Reproductive Health, University of Oxford, Oxford, UK
Pliable rejection sampling
Apr 24, 2026Akram Erraqabi, Michal Valko, Alexandra Carpentier +1Rejection Sampling
MILA, Universit´e de Montr´eal, Montr´eal, QC H3C 3J7, Canada · INRIA Lille - Nord Europe, SequeL team, 40 avenue Halley 59650, Villeneuve d’Ascq, France · Institut für Mathematik, Univertät Potsdam, Germany, Haus 9 Karl-Liebknecht-Strasse 24-25, D-14476 Potsdam +1
On the Rejection Criterion for Proxy-based Test-time Alignment
Apr 17, 2026Ayoub Hammal, Pierre Zweigenbaum, Caio CorroLLM Alignment Inference-Time Language Model Alignment
Université Paris-Saclay, CNRS, LISN · INSA Rennes, IRISA, CNRS, Université de Rennes
Video Replanning via Latent Embedding Refinement and Rejection-Based Sampling
Oct 20, 2025Po-Chen Ko, Yu Chan, Yu-Hsiang Fu +7Robot System Identification Robot Failure Recovery
National Taiwan University · MIT CSAIL · Cornell University +1

Oct 3, 2026J/K move · Enter open · S save

Verdicts Without Annotated Evidence: Rejection Sampling or Label-Only Post-Training for Evidence Recovery?

Nishanth Nayakanti, Prasang Gupta, Ashutosh Bilthare, Kevin Paul

PricewaterhouseCoopers, U.S.

Language Model Post-Training Rejection Sampling Natural Language Inference

Open paper AskPDF arXiv

In many review workflows the verdict is the only thing retained. The passages behind it are not marked, because that annotation costs far more than recording the decision. We measure how much of that evidence a small language model can recover when it is post-trained on the verdicts alone, with no human evidence labels at any stage. On ContractNLI the human evidence spans are held out until evaluation. Matching the recorded verdict and agreeing with those spans are not the same thing: across six systems the two scores are only weakly related and rank the systems differently, so accuracy is a poor guide when the citations have to be reviewable. Label-only training on the bare verdict reaches accuracy 0.896 and span F1 0.564. Rejection sampling, which keeps a generated trace only when its verdict matches the record and then picks one by an automatic source-grounding score, reaches 0.797 and 0.556, against 0.747 and 0.493 before training. Verbatim citation rises from 0.597 to 0.729 under label-only training and to 0.701 under rejection sampling. One seed on one corpus cannot say which method is better, but both improve the evidence without anyone annotating it.