Discrete Tilt Matching

Researchers have developed Discrete Tilt Matching (DTM), a likelihood-free method for fine-tuning masked diffusion large language models using reinforcement learning. The approach recasts fine-tuning as state-level matching of local unmasking posteriors under reward tilting, enabling training stability through control variates and an annealing schedule. When applied to the LLaDA-8B-Instruct model, DTM achieved strong performance gains on Sudoku and Countdown tasks while remaining competitive on MATH500 and GSM8K benchmarks.

Computer Science Machine Learning Submitted on 20 Apr 2026 v1 https://arxiv.org/abs/2604.18739v1 , last revised 19 May 2026 this version, v3 Title:Discrete Tilt Matching View PDF /pdf/2604.18739 HTML experimental https://arxiv.org/html/2604.18739v3 Abstract:Masked diffusion large language models dLLMs are a promising alternative to autoregressive generation. While reinforcement learning RL methods have recently been adapted to dLLM fine-tuning, their objectives typically depend on sequence-level marginal likelihoods, which are intractable for masked diffusion models. To address this, we derive Discrete Tilt Matching DTM , a likelihood-free method that recasts dLLM fine-tuning as state-level matching of local unmasking posteriors under reward tilting. DTM takes the form of a weighted cross-entropy objective with explicit minimizer, and admits control variates that improve training stability. On a synthetic maze-planning task, we analyze how DTM's annealing schedule and control variates affect training stability and prevent mode collapse. At scale, fine-tuning LLaDA-8B-Instruct with DTM yields strong gains on Sudoku and Countdown while remaining competitive on MATH500 and GSM8K. Submission history From: Yuyuan Chen view email /show-email/a0a9fb9a/2604.18739 Mon, 20 Apr 2026 18:43:37 UTC 6,659 KB v1 /abs/2604.18739v1 Sat, 16 May 2026 03:48:01 UTC 6,659 KB v2 /abs/2604.18739v2 v3 Tue, 19 May 2026 01:39:31 UTC 6,659 KB Current browse context: cs.LG References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender IArxiv Recommender What is IArxiv? https://iarxiv.org/about arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .