Answer First, Reason Later: When Commitment Order Costs Accuracy in Diffusion Language Models
Jewon Yeom, Jaewon Sok, Seonghyeon Park, Jeongjae Park, Hwiyeong Lee, Taesup Kim
cs.CL, cs.AI
Submitted: 2026-08-24
Updated: 2026-08-25
Code: https://github.com/ML-GSAI/LLaDA
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models
- Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought
- Training Verifiers to Solve Math Word Problems
- Robust Answers, Fragile Logic: Probing the Decoupling Hypothesis in LLM Reasoning
- "Well, Keep Thinking": Enhancing LLM Reasoning with Adaptive Injection Decoding
- Rainbow Padding: Mitigating Early Termination in Instruction-Tuned Diffusion LLMs
- Measuring Faithfulness in Chain-of-Thought Reasoning
- VoidPadding: Let [VOID] Handle Padding in Masked Diffusion Language Models so that [EOS] Can Focus on Semantic Termination
- Measuring Temporal Linguistic Emergence in Diffusion Language Models
- The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models
- Large Language Diffusion Models
- Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning
- Deferred Commitment Decoding for Diffusion Language Models
- DLM-Scope: Mechanistic Interpretability of Diffusion Language Models via Sparse Autoencoders
- On the Trainability of Masked Diffusion Language Models via Blockwise Locality
- $\rho$-$\texttt{EOS}$: Training-free Bidirectional Variable-Length Control for Masked Diffusion LLMs
- MMaDA: Multimodal Large Diffusion Language Models
- Diffusion LLM with Native Variable Generation Lengths: Let [EOS] Lead the Way
- Decoding in Order-Agnostic Language Models: Chain-Rule Deviation and Uniform Spreading
- Dream 7B: Diffusion Large Language Models
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering