mlsanityData to 7 Oct 2026

Learnability-Informed Fine-Tuning of Diffusion Language Models

arXiv 2605.22939 ↗First posted 21 May 2026cs.CL cs.LGcited by 3

On the map: LLM reasoning → Diffusion reasoning and test-time scaling

also in 1 more
Abstract (arXiv)

We aim to improve the reasoning capabilities of diffusion language models (DLMs). While SFT is a popular post-training recipe for autoregressive models, its use in DLMs faces challenges and can even hurt performance, though the underlying causes remain understudied. Our analysis reveals that vanilla SFT overlooks learnability, namely what and when tokens are learned. Specifically, rare tokens are difficult to learn when most of the input is masked, whereas it is straightforward and thus of little value to learn common tokens when most of the input is unmasked. Motivated by our analysis, we propose LIFT, an efficient SFT-based post-training algorithm for DLMs. LIFT learns easy tokens when most of the input is masked and hard tokens when more context is available, thus aligning the training with the information available at different diffusion time steps. Our results show that LIFT outperforms existing SFT baselines across six reasoning benchmarks, achieving up to a 3x relative gain on AIME'24 and AIME'25. Our code is publicly available at https://github.com/divelab/LIFT.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Placed in Diffusion reasoning and test-time scaling, one of 17 directions in LLM reasoning. Builds on 32 of its 38 references that are in the map. First posted on arXiv 21 May 2026.

Builds on

32 of 38 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025d2: Improving Reasoning in Diffusion Language Models via Trajectory Likelihood Estimation25 Sep
  2. 2025Mercury: Ultra-Fast Language Models Based on Diffusion17 Jun
  3. 2025d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning16 Apr

Other references · 29

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  4. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  6. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  7. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
  8. 2021Program Synthesis with Large Language Modelscited by 3,509
  9. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
  10. 2022Video Diffusion Modelscited by 2,280
  11. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
  12. 2025s1: Simple test-time scalingcited by 1,311
17 more
  1. 2021Variational Diffusion Modelscited by 1,145
  2. 2025Large Language Diffusion Modelscited by 872
  3. 2024Simple and Effective Masked Diffusion Language Modelscited by 800
  4. 2025Dream 7B: Diffusion Large Language Modelscited by 515
  5. 2022Large Language Models Struggle to Learn Long-Tail Knowledgecited by 452
  6. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 404
  7. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 273
  8. 2025LLaDA2.0: Scaling Up Diffusion Language Models to 100Bcited by 154
  9. 2025Llama-Nemotron: Efficient Reasoning Modelscited by 151
  10. 2024Stream of Search (SoS): Learning to Search in Languagecited by 141
  11. 2025KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Codingcited by 115
  12. 2025Self-Evolving Curriculum for LLM Reasoningcited by 107
  13. 2025Curriculum Reinforcement Learning from Easy to Hard Tasks Improves LLM Reasoningcited by 89
  14. 2024No "Zero-Shot" Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performancecited by 75
  15. 2024The Neglected Tails in Vision-Language Modelscited by 63
  16. 2025GIFT: Guided Importance-Aware Fine-Tuning for Diffusion Language Modelscited by 4
  17. 2026Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantagescited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026CanvasAnneal: Curriculum Reinforcement Learning for Diffusion Language Modelscited by 0
  2. 2026Fine-Tuning Diffusion Language Models with Context Selection and Target Weightingcited by 0
  3. 2026Enhancing Diffusion Language Models with Autoregressive Post-Training Weightscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Reasoning with Latent Tokens in Diffusion Language Models
  2. 2025TESS 2: A Large-Scale Generalist Diffusion Language Model
  3. 2025A Survey on Diffusion Language Models
  4. 2025Diffusion Language Models are Super Data Learners
  5. 2026Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models
  6. 2026Diffusion Language Models: An Experimental Analysis
  7. 2025Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner
  8. 2026Towards Efficient Reasoning: Learning Causal Shortcuts for Diffusion Language Models