AI News HubLIVE
Original source2 min read

Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models

Diffusion language models are a promising alternative to autoregressive models, but current post-training methods that maximize rewards lead to a failure mode called trajectory locking, where probability mass concentrates on a narrow set of denoising paths, reducing coverage of alternative correct solutions. The proposed TraFL (Trajectory Flow baLancing) uses a trajectory-balance objective to train the policy toward a reward-tilted target distribution anchored to a frozen reference model, with a diffusion-compatible sequence-level surrogate and learned prompt-dependent normalization. TraFL is the only method that improves over the base model in all benchmark settings, with gains that persist as sampling budget increases, and transfers to held-out evaluations.

SourcearXiv Machine LearningAuthor: Saba Ahmadi, Prasanna Parthasarathi, Yufei Cui

[2605.13935] Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models

[Submitted on 13 May 2026]

Title:Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models

View a PDF of the paper titled Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models, by Saba Ahmadi and 2 other authors

View PDF HTML (experimental)

Abstract:Diffusion language models are a promising alternative to autoregressive models, yet post-training methods for them largely adapt reward-maximizing objectives. We identify a central failure mode in this setting we call trajectory locking: sampled reward-driven updates over-concentrate probability mass onto a narrow set of denoising paths, reducing coverage of alternative correct solutions under repeated sampling. To address this, we propose TraFL (Trajectory Flow baLancing), a trajectory-balance objective that trains the policy toward a reward-tilted target distribution anchored to a frozen reference model. We make this practical for diffusion language models with a diffusion-compatible sequence-level surrogate and a learned prompt-dependent normalization. Across mathematical reasoning and code generation benchmarks, TraFL is the only evaluated post-training method that improves over the base model in every benchmark-length setting, with gains that persist as the sampling budget increases. The improvements transfer to held-out evaluations: TraFL stays above the base model on Minerva Math and is the strongest method on every LiveCodeBench difficulty split.

Subjects:

Machine Learning (cs.LG); Computation and Language (cs.CL)

Cite as: arXiv:2605.13935 [cs.LG]

(or arXiv:2605.13935v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2605.13935

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Saba Ahmadi [view email] [v1] Wed, 13 May 2026 16:14:46 UTC (107 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models, by Saba Ahmadi and 2 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.LG

new | recent | 2026-05

Change to browse by:

cs cs.CL

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

IArxiv recommender toggle

IArxiv Recommender (What is IArxiv?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)