Papers with rejection sampling
On Sample-Efficient Code Generation (2023.emnlp-industry)
Copied to clipboard
Hojae Han, Yu Jin Kim, Byoungjip Kim, Youngwon Lee, Kyungjae Lee, Kyungmin Lee, Moontae Lee, Kyunghoon Bae, Seung-won Hwang
| Challenge: | Existing approaches to code generation rely on rejection sampling to generate multiple code snippets then select the best. |
| Approach: | They propose a framework that prioritizes sampling on test problems that models can solve. |
| Outcome: | The proposed framework reduces sampling costs while maintaining comparable code generation performance. |
RS-DPO: A Hybrid Rejection Sampling and Direct Preference Optimization Method for Alignment of Large Language Models (2024.findings-naacl)
Copied to clipboard
| Challenge: | Reinforcement learning with human feedback (RLHF) is widely employed to align large language models with user intent. |
| Approach: | They propose to combine rejection sampling and direct preference optimization to improve alignment with user intent by identifying pairs of contrastive samples from human annotator and alternative LLMs. |
| Outcome: | The proposed method outperforms existing methods including RS, PPO, and DPO in a limited resource environment. |
Think-Search-Patch: A Retrieval-Augmented Reasoning Framework for Repository-Level Code Repair (2025.emnlp-industry)
Copied to clipboard
Bojian Xiong, Yikun Lei, Xikai Liu, Shaowei Zhang, Pengyun Zhu, Yan Liu, Yongqi Leng, Ling Shi, Meizhi Zhong, Yurong Zhang, Yan Gao, null Yiwu, Yao Hu, Deyi Xiong
| Challenge: | Large language models suffer from multiple-file coding scenarios with strong inter-file dependencies . experimental results show that large language models exhibit inadequate performance in multi-file scenarios . |
| Approach: | They propose a retrieval-augmented reasoning framework for repository-level code repair . they use a dataset to generate standardized patches based on the key snippets . |
| Outcome: | The proposed framework improves retrieval accuracy and repair success on SWE-bench Lite . it surpasses models with larger size in managing extensive code contexts and fixing bugs spanning across multiple files. |
LegalDrill: Diagnosis-Driven Synthesis for Legal Reasoning in Small Language Models (2026.acl-industry)
Copied to clipboard
Tianchun Li, Haochen Liu, Vishwa Pardeshi, Xingchen Wang, Tianci Liu, Huijun Zhao, Wei Fan, Jing Gao
| Challenge: | Small language models (SLMs) are promising for real-world deployment but struggle with high-stakes legal reasoning tasks. |
| Approach: | They propose a diagnostic-driven synthesis framework that extracts and refines reasoning trajectories from a capable teacher via fine-grained prompting and a self-reflective verification is employed to adaptively select the most effective data for the SLM student. |
| Outcome: | The proposed framework extracts and refines reasoning trajectories from a capable teacher via fine-grained prompting, then a self-reflective verification is employed to adaptively select the most effective data for the student. |
SemStamp: A Semantic Watermark with Paraphrastic Robustness for Text Generation (2024.naacl-long)
Copied to clipboard
Abe Hou, Jingyu Zhang, Tianxing He, Yichen Wang, Yung-Sung Chuang, Hongwei Wang, Lingfeng Shen, Benjamin Van Durme, Daniel Khashabi, Yulia Tsvetkov
| Challenge: | Existing watermarked generation algorithms employ token-level designs and are vulnerable to paraphrase attacks. |
| Approach: | They propose a sentence-level watermarking algorithm that uses locality-sensitive hashing to partition the semantic space of sentences. |
| Outcome: | The proposed algorithm is more robust than the existing state-of-the-art method on paraphrasers and domains, while posing only minor degradations to SemStamp. |
Crowd Comparative Reasoning: Unlocking Comprehensive Evaluations for LLM-as-a-Judge (2025.acl-long)
Copied to clipboard
Qiyuan Zhang, Yufei Wang, Yuxin Jiang, Liangyou Li, Chuhan Wu, Yasheng Wang, Xin Jiang, Lifeng Shang, Ruiming Tang, Fuyuan Lyu, Chen Ma
| Challenge: | Existing methods rely on majority voting or criteria expansion to capture detailed and detailed details, often leading to incomplete outcomes. |
| Approach: | They propose a method which introduces additional crowd responses to compare with the candidate responses, thereby exposing deeper and more comprehensive details within the candidate answers. |
| Outcome: | Experiments show that the proposed method improves evaluation reliability and achieves an average gain of 6.7% across five benchmarks. |
Beyond Rejection Sampling: Trajectory Fusion for Scaling Mathematical Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) fine-tuned using rejection sampling retain only correct reasoning trajectories . however, this paradigm treats supervision as a binary filter that systematically excludes teacher-generated errors, leaving a gap in how reasoning failures are modeled during training. |
| Approach: | They propose a fine-tuning strategy that reframes rejection sampling as a structured supervision construction process. |
| Outcome: | The proposed approach outperforms RFT on multiple math benchmarks while retaining only correct reasoning trajectories. |
Debt Collection Negotiations with Large Language Models: An Evaluation System and Optimizing Decision Making with Multi-Agent (2025.findings-acl)
Copied to clipboard
| Challenge: | Debt collection negotiations (DCN) are vital for managing non-performing loans (NPLs) prior systems lacking dynamic negotiation and real-time decision-making capabilities. |
| Approach: | They propose a framework for debt negotiation that incorporates planning and judging modules to improve decision rationality. |
| Outcome: | The proposed framework improves decision rationality and integrates planning and judging modules to improve decision rationalness. |
GALA: Geometric Data Selection with Strategic Prospecting for Large Language Model Self-training (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to self-training are based on reject sampling and lack quality reasoning paths. |
| Approach: | They propose a framework for self-training using a generate-and-filter paradigm . they propose to identify diverse and informative samples from redundant data and exploit them more strategically. |
| Outcome: | The proposed framework exploits informative samples from redundant data and improves reasoning trajectory prospecting. |
Learning from Contrasts: Synthesizing Reasoning Paths from Diverse Search Trajectories (2026.acl-long)
Copied to clipboard
| Challenge: | MCTS methods retain only the single highest-reward trajectory, discarding comparative signals present in the many explored paths. |
| Approach: | They propose a framework that transforms supervision extraction into a synthesis procedure. |
| Outcome: | The proposed framework matches or exceeds baselines on 60K CRPS-synthesized examples on out-of-domain benchmarks. |
A Systematic Examination of Preference Learning through the Lens of Instruction-Following (2025.naacl-long)
Copied to clipboard
Joongwon Kim, Anirudh Goyal, Aston Zhang, Bo Xiong, Rui Hou, Melanie Kambadur, Dhruv Mahajan, Hannaneh Hajishirzi, Liang Tan
| Challenge: | a recent study has found that preference learning is a key tool for enhancing LLM training and alignment. |
| Approach: | They use a synthetic data generation pipeline to generate 48,000 unique instruction-following prompts with 23 verifiable constraints to obtain preference pairs. |
| Outcome: | The proposed pipeline generates 48,000 unique instruction-following prompts with 23 verifiable constraints that enable fine-grained and automated quality assessments of model responses. |
Learning Rhyming Constraints using Structured Adversaries (D19-1)
Copied to clipboard
| Challenge: | Existing approaches to text generation fail to capture higher-level structure in text, for example, rhyming patterns. |
| Approach: | They propose a method that uses a structured discriminator to learn rhyming constraints from poetry . the discriminator compares two English poetry datasets based on a learned similarity matrix . |
| Outcome: | The proposed method can learn rhyming patterns in English poetry without explicit phonetic information. |
Correct, Concise and Complete: Multi-stage Training For Adaptive Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) increase test-time computation, often in the form of chain-of-thought (CoT) however, reasoning traces can become unnecessarily long, increasing computation costs without improving accuracy and sometimes even degrading performance. |
| Approach: | They propose a multi-stage efficient reasoning method that combines supervised fine-tuning with reinforcement learning using an adaptive length penalty. |
| Outcome: | The proposed method reduces response length by an average of 28% for 8B models and 40% for 32B models while incurring only minor performance drops of 1.6 and 2.5 points, respectively. |
RoseRAG: Robust Retrieval-augmented Generation with Small-scale LLMs via Margin-aware Preference Optimization (2025.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) have impressive performance but face high computational costs and latency, limiting their deployment in resource-constrained settings. |
| Approach: | They propose a robust RAG framework for large language models via Margin-aware Preference Optimization to enhance the accuracy and reliability of SLMs. |
| Outcome: | The proposed framework surpasses state-of-the-art benchmarks on three open-domain question answering tasks. |
MM-Verify: Enhancing Multimodal Reasoning with Chain-of-Thought Verification (2025.acl-long)
Copied to clipboard
| Challenge: | MM-Verifier and MM Reasoner are a powerful multimodal reasoning model . large language models (LLMs) have demonstrated exceptional performance across tasks spanning myriad domains. |
| Approach: | They propose a method which combines tree search and verification to generate high-quality chain-of-thought data. |
| Outcome: | The proposed method outperforms all larger models on the MathCheck, MathVista, and MathVerse benchmarks. |
Learning New Skills after Deployment: Improving open-domain internet-driven dialogue with human feedback (2023.acl-long)
Copied to clipboard
| Challenge: | Frozen models trained to mimic static datasets can never improve their performance. |
| Approach: | They propose to use binary quality measurements and free-form text feedback to improve conversational skills in a conversational learning framework. |
| Outcome: | The proposed model improves on the DIRECTOR model, which is based on binary quality measurements and free-form text feedback, and shows that iterative retraining and redeployment can improve the model. |
When LLMs Read Tables Carelessly: Measuring and Reducing Data Referencing Errors (2026.acl-long)
Copied to clipboard
Yuqing Yang, Qi Zhu, Zhen Han, Boran Han, Zhengyuan Shen, Shuai Wang, Vassilis N. Ioannidis, Huzefa Rangwala
| Challenge: | Large language models (LLMs) perform well on table tasks, but they still make data referencing errors (DREs) prior studies have only offered limited, small-scale analyses. |
| Approach: | They propose inference-time strategies and lightweight critics to mitigate data referencing errors. |
| Outcome: | The proposed model achieves an average F1 score of 78.2% in detecting both in-distribution and out-of-difference DREs and assists inference for larger models. |
Multi-modal Preference Alignment Remedies Degradation of Visual Instruction Tuning on Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | Multi-modal large language models (MLLMs) are expected to support multi-turn queries of interchanging image and text modalities in production. |
| Approach: | They propose to use visual-question-answering (VQA) datasets to annotate a 5k-sample VQA preference dataset and to investigate the degradation of VQA datasets. |
| Outcome: | The proposed model surpasses the instruction-following capabilities of the language model with DPO and SteerLM. |
GRaMPa: Subword Regularisation by Skewing Uniform Segmentation Distributions with an Efficient Path-counting Markov Model (2025.acl-long)
Copied to clipboard
| Challenge: | Subword regularisations are known to be stochastic, but only a handful of possible segmentations are sampled. |
| Approach: | They propose to randomise word segmentations from a subword tokeniser instead of randomising them by weighting paths in an unweighted segmentation graph. |
| Outcome: | The proposed method outperforms existing methods on token-level tasks with spelling errors. |
FastMCTS: A Simple Sampling Strategy for Data Synthesis (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for generating multi-step reasoning data rely on rejection sampling, which generates trajectories independently and suffers from inefficiency and imbalanced sampling across problems of varying difficulty levels. |
| Approach: | They propose a data synthesis strategy inspired by Monte Carlo Tree Search . it offers step-level evaluation signals and promotes balanced sampling . |
| Outcome: | Experiments show that FastMCTS generates 30% more correct reasoning paths than rejection sampling. |
ExeSQL: Self-Taught Text-to-SQL Models with Execution-Driven Bootstrapping for SQL Dialects (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing text-to-SQL models are limited to SQLite due to dataset limitations . data generated through static prompting is noisy and unreliable, authors say . |
| Approach: | They propose a text-to-SQL framework with execution-driven, agentic bootstrapping . ExeSQl bridges the dialect gap in text- to-Sql, achieving average improvements . |
| Outcome: | ExeSQL bridges the dialect gap in text-to-SQl, with average improvements of 15.2%, 10.38%, and 4.49% over GPT-4o on PostgreSQLE, MySQL, and Oracle. |
SimMark: A Robust Sentence-Level Similarity-Based Watermarking Algorithm for Large Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Experimental results demonstrate that SimMark surpasses previous sentence-level watermarking techniques in robustness, sampling efficiency, and applicability across diverse domains. |
| Approach: | They propose a sentence-level watermarking algorithm that makes LLM outputs traceable without requiring access to model internals. |
| Outcome: | The proposed algorithm surpasses previous sentence-level watermarking techniques in robustness, sampling efficiency, and applicability across diverse domains while maintaining the text quality and fluency. |