Fork-dLLM: Avoiding the Flexibility Trap in Diffusion Language Models
cs.LG
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/LeapLabTHU/JustGRPO
Sources
- Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models
- Program Synthesis with Large Language Models
- The Design Space of Tri-Modal Masked Diffusion Models
- LLaDA2.0: Scaling Up Diffusion Language Models to 100B
- Evaluating Large Language Models Trained on Code
- dUltra: Ultra-Fast Diffusion Language Models via Reinforcement Learning
- DMax: Aggressive Parallel Decoding for dLLMs
- Training Verifiers to Solve Math Word Problems
- From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models
- Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages
- Free Lunch for Pass@$k$? Low Cost Diverse Sampling for Diffusion Language Models
- DACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models
- The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models
- Improved Large Language Diffusion Models
- Mask-Aware Policy Gradients for Diffusion Language Models
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Improving Diffusion Language Model Decoding through Joint Search in Generation Order and Token Space
- GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models
- DiffusionGemma Technical Report
- DUEL: Exact Likelihood for Masked Diffusion via Deterministic Unmasking
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks