Alpha Diffusion Language Models: Factorization Alone Is Not the Problem

arXiv:2609.38066 · cs.LG · Submitted 2026-09-29 · Read on arXiv

cs.LG

Submitted: 2026-09-29

Updated: 2026-09-29

Project page: https://gaotangli.github.io/project_page/Beyond-Log-Likelihood/static/pdfs/beyond-log-likelihood.pdf

Terminology

• alpha — 1× in this paper · explained in 1 episode

Sources

• Posterior Refinement: Fast Language Generation via Any-Order Flow Maps
• OpenCodeInstruct: A Large-scale Instruction Tuning Dataset for Code LLMs
• Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models
• Structured Denoising Diffusion Models in Discrete State-Spaces
• Program Synthesis with Large Language Models
• Enabling Approximate Joint Sampling in Diffusion LMs
• LLaDA2.0: Scaling Up Diffusion Language Models to 100B
• Evaluating Large Language Models Trained on Code
• Language Modeling with Hyperspherical Flows
• Non-Autoregressive Neural Machine Translation
• Measuring Mathematical Problem Solving With the MATH Dataset
• Tensor-Train Joint Modeling for Few-Step Discrete Diffusion
• CDLM: Consistency Diffusion Language Models For Faster Sampling
• IDLM: Inverse-distilled Diffusion Language Models
• TinyGSM: achieving >80% on GSM8k with small language models
• Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution
• Simple and Effective Masked Diffusion Language Models
• The Diffusion Duality
• Simplified and Generalized Masked Diffusion for Discrete Data
• A Tunable Loss Function for Robust Classification: Calibration, Landscape, and Generalization

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