Image de couverture de livre

Don't Retrain, Align: Adapting Autoregressive LMs to Diffusion LMs via Representation Alignment

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Peng, Fred Zhangzhi, Fox, Alexis, Zhang, Anru R., Tong, Alexander
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

DiffuMamba: High-Throughput Diffusion LMs with Mamba Backbone
par: Singh, Vaibhav, et autres
Publié: (2025)

Evaluating Generalization and Representation Stability in Small LMs via Prompting, Fine-Tuning and Out-of-Distribution Prompts
par: Raja, Rahul, et autres
Publié: (2025)

Don't Retrain, Just Reuse: Recovering Dual-Target Molecules from Single-Target Diffusion Models
par: Zeng, Qingyuan, et autres
Publié: (2026)

Don't Blind Your VLA: Aligning Visual Representations for OOD Generalization
par: Kachaev, Nikita, et autres
Publié: (2025)

Grounded Token Initialization for New Vocabulary in LMs for Generative Recommendation
par: Chen, Daiwei, et autres
Publié: (2026)

Coupling Models for One-Step Discrete Generation
par: Peng, Fred Zhangzhi, et autres
Publié: (2026)

Beyond Binary Rewards: Training LMs to Reason About Their Uncertainty
par: Damani, Mehul, et autres
Publié: (2025)

Path Planning for Masked Diffusion Model Sampling
par: Peng, Fred Zhangzhi, et autres
Publié: (2025)

Forecasting Application Counts in Talent Acquisition Platforms: Harnessing Multimodal Signals using LMs
par: Kabir, Md Ahsanul, et autres
Publié: (2024)

When Thoughts Meet Facts: Reusable Reasoning for Long-Context LMs
par: Jeong, Soyeong, et autres
Publié: (2025)

Addressing the Ecological Fallacy in Larger LMs with Human Context
par: Soni, Nikita, et autres
Publié: (2026)

OpenScholar: Synthesizing Scientific Literature with Retrieval-augmented LMs
par: Asai, Akari, et autres
Publié: (2024)

Safety Alignment of LMs via Non-cooperative Games
par: Paulus, Anselm, et autres
Publié: (2025)

Dodo: Dynamic Contextual Compression for Decoder-only LMs
par: Qin, Guanghui, et autres
Publié: (2023)

Don't Freeze, Don't Crash: Extending the Safe Operating Range of Neural Navigation in Dense Crowds
par: Zhang, Jiefu, et autres
Publié: (2026)

ImProver 2: Iteratively Self-Improving LMs for Neurosymbolic Proof Optimization
par: Ahuja, Riyaz, et autres
Publié: (2026)

$\textbf{PLUM}$: Improving Code LMs with Execution-Guided On-Policy Preference Learning Driven By Synthetic Test Cases
par: Zhang, Dylan, et autres
Publié: (2024)

Don't Start from Scratch: Behavioral Refinement via Interpolant-based Policy Diffusion
par: Chen, Kaiqi, et autres
Publié: (2024)

SentinelLMs: Encrypted Input Adaptation and Fine-tuning of Language Models for Private and Secure Inference
par: Mishra, Abhijit, et autres
Publié: (2023)

NoFunEval: Funny How Code LMs Falter on Requirements Beyond Functional Correctness
par: Singhal, Manav, et autres
Publié: (2024)

Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior Prediction
par: Rector-Brooks, Jarrid, et autres
Publié: (2024)

Don't Judge by the Look: Towards Motion Coherent Video Representation
par: Zhang, Yitian, et autres
Publié: (2024)

Entropy-Aligned Decoding of LMs for Better Writing and Reasoning
par: Ahmed, Kareem, et autres
Publié: (2026)

Planner Aware Path Learning in Diffusion Language Models Training
par: Peng, Fred Zhangzhi, et autres
Publié: (2025)

Interpretable-by-Design Transformers via Architectural Stream Independence
par: Kerce, Clayton, et autres
Publié: (2026)

Transformers Don't In-Context Learn Least Squares Regression
par: Hill, Joshua, et autres
Publié: (2025)

Don't Play Favorites: Minority Guidance for Diffusion Models
par: Um, Soobin, et autres
Publié: (2023)

Enabling Approximate Joint Sampling in Diffusion LMs
par: Bansal, Parikshit, et autres
Publié: (2025)

Don't be lazy: CompleteP enables compute-efficient deep transformers
par: Dey, Nolan, et autres
Publié: (2025)

Improving Policy Optimization via $\varepsilon$-Retrain
par: Marzari, Luca, et autres
Publié: (2024)

Benchmarking is Broken -- Don't Let AI be its Own Judge
par: Cheng, Zerui, et autres
Publié: (2025)

Don't Forget It! Conditional Sparse Autoencoder Clamping Works for Unlearning
par: Khoriaty, Matthew, et autres
Publié: (2025)

Do's and Don'ts: Learning Desirable Skills with Instruction Videos
par: Kim, Hyunseung, et autres
Publié: (2024)

Don't Waste Your Time: Early Stopping Cross-Validation
par: Bergman, Edward, et autres
Publié: (2024)

Don't Trust: Verify -- Grounding LLM Quantitative Reasoning with Autoformalization
par: Zhou, Jin Peng, et autres
Publié: (2024)

xAI-Drop: Don't Use What You Cannot Explain
par: De Luca, Vincenzo Marco, et autres
Publié: (2024)

Don't Lag, RAG: Training-Free Adversarial Detection Using RAG
par: Kazoom, Roie, et autres
Publié: (2025)

Position: Don't Use the CLT in LLM Evals With Fewer Than a Few Hundred Datapoints
par: Bowyer, Sam, et autres
Publié: (2025)

Linear Transformers as VAR Models: Aligning Autoregressive Attention Mechanisms with Autoregressive Forecasting
par: Lu, Jiecheng, et autres
Publié: (2025)

Reasoning Models Don't Always Say What They Think
par: Chen, Yanda, et autres
Publié: (2025)