AR-MAP: Are Autoregressive Large Language Models Implicit Teachers for Diffusion Large Language Models?

• URL: http://arxiv.org/abs/2602.02178v2
• Date: Tue, 03 Feb 2026 03:22:33 GMT
• Title: AR-MAP: Are Autoregressive Large Language Models Implicit Teachers for Diffusion Large Language Models?
• Authors: Liang Lin, Feng Xiong, Zengbin Wang, Kun Wang, Junhao Dong, Xuecai Hu, Yong Wang, Xiangxiang Chu,
• Abstract summary: Diffusion Large Language Models (DLLMs) have emerged as a powerful alternative to autoregressive models.<n> preference alignment ofDLLMs remains challenging due to high variance introduced by Evidence Lower Bound (ELBO)-based likelihood estimation.<n>We propose AR-MAP, a novel transfer learning framework that leverages preference-aligned autoregressive LLMs as implicit teachers for divergentM alignment.
• Score: 58.52365018076441
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Diffusion Large Language Models (DLLMs) have emerged as a powerful alternative to autoregressive models, enabling parallel token generation across multiple positions. However, preference alignment of DLLMs remains challenging due to high variance introduced by Evidence Lower Bound (ELBO)-based likelihood estimation. In this work, we propose AR-MAP, a novel transfer learning framework that leverages preference-aligned autoregressive LLMs (AR-LLMs) as implicit teachers for DLLM alignment. We reveal that DLLMs can effectively absorb alignment knowledge from AR-LLMs through simple weight scaling, exploiting the shared architectural structure between these divergent generation paradigms. Crucially, our approach circumvents the high variance and computational overhead of direct DLLM alignment and comprehensive experiments across diverse preference alignment tasks demonstrate that AR-MAP achieves competitive or superior performance compared to existing DLLM-specific alignment methods, achieving 69.08\% average score across all tasks and models. Our Code is available at https://github.com/AMAP-ML/AR-MAP.

Related papers

• Explore More, Learn Better: Parallel MLLM Embeddings under Mutual Information Minimization [35.43577499735611]
  We introduce one Parallel Decoupling Framework (PDF) for multimodal embedding learning.<n>PDF conditions a shared MLLM backbone on distinct, learnable prefixes to roll out multiple parallel paths for one input.<n>We instantiate PDF on multiple MLLM backbones and prove its effectiveness on MMEB benchmark.
  arXiv  Detail & Related papers  (2025-11-03T13:57:08Z)
• OptMerge: Unifying Multimodal LLM Capabilities and Modalities via Model Merging [124.91183814854126]
  Model merging seeks to combine multiple expert models into a single model.<n>We introduce a benchmark for model merging research that clearly divides the tasks for MLLM training and evaluation.<n>We find that model merging offers a promising way for building improved MLLMs without requiring training data.
  arXiv  Detail & Related papers  (2025-05-26T12:23:14Z)
• Your Mixture-of-Experts LLM Is Secretly an Embedding Model For Free [21.59456761618456]
  Large language models (LLMs) excel on generation tasks, their decoder-only architecture often limits their potential as embedding models if no further representation finetuning is applied. Our study shows that the expert routers in MoE LLMs can serve as an off-the-shelf embedding model with promising performance on a diverse class of embedding-focused tasks.
  arXiv  Detail & Related papers  (2024-10-14T17:59:44Z)
• Align$^2$LLaVA: Cascaded Human and Large Language Model Preference Alignment for Multi-modal Instruction Curation [56.75665429851673]
  This paper introduces a novel instruction curation algorithm, derived from two unique perspectives, human and LLM preference alignment.<n>Experiments demonstrate that we can maintain or even improve model performance by compressing synthetic multimodal instructions by up to 90%.
  arXiv  Detail & Related papers  (2024-09-27T08:20:59Z)
• ULLME: A Unified Framework for Large Language Model Embeddings with Generation-Augmented Learning [72.90823351726374]
  We introduce the Unified framework for Large Language Model Embedding (ULLME), a flexible, plug-and-play implementation that enables bidirectional attention across various LLMs. We also propose Generation-augmented Representation Learning (GRL), a novel fine-tuning method to boost LLMs for text embedding tasks. To showcase our framework's flexibility and effectiveness, we release three pre-trained models from ULLME with different backbone architectures.
  arXiv  Detail & Related papers  (2024-08-06T18:53:54Z)
• Understanding Alignment in Multimodal LLMs: A Comprehensive Study [46.33812471516309]
  We analyze each aspect of preference alignment in Multimodal Large Language Models (MLLMs) We show that combining offline and online methods can improve the performance of the model in certain scenarios. We introduce a novel way of creating multimodal preference data called Bias-Driven Hallucination Sampling (BDHS)
  arXiv  Detail & Related papers  (2024-07-02T17:55:03Z)
• Large Language Models Can Automatically Engineer Features for Few-Shot Tabular Learning [35.03338699349037]
  We propose a novel in-context learning framework, FeatLLM, which employs Large Language Models as feature engineers. FeatLLM generates high-quality rules, significantly (10% on average) outperforming alternatives such as TabLLM and STUNT.
  arXiv  Detail & Related papers  (2024-04-15T06:26:08Z)
• Knowledge Fusion of Large Language Models [73.28202188100646]
  This paper introduces the notion of knowledge fusion for large language models (LLMs) We externalize their collective knowledge and unique strengths, thereby elevating the capabilities of the target model beyond those of any individual source LLM. Our findings confirm that the fusion of LLMs can improve the performance of the target model across a range of capabilities such as reasoning, commonsense, and code generation.
  arXiv  Detail & Related papers  (2024-01-19T05:02:46Z)
• LLM-Pruner: On the Structural Pruning of Large Language Models [65.02607075556742]
  Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. We tackle the compression of LLMs within the bound of two constraints: being task-agnostic and minimizing the reliance on the original training dataset. Our method, named LLM-Pruner, adopts structural pruning that selectively removes non-critical coupled structures.
  arXiv  Detail & Related papers  (2023-05-19T12:10:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.