Self-conditioned flow language models solve fixed-point iterations, enabling fixed-point flow maps that distill into FMLM* which outperforms SOTA in few-step generation on OpenWebText.
hub
LangFlow: Continuous Diffusion Rivals Discrete in Language Modeling
14 Pith papers cite this work. Polarity classification is still indexing.
14
Pith papers citing it
abstract
Continuous diffusion has been the foundation of high-fidelity, controllable, and few-step generation of many data modalities such as images. However, in language modeling, prior continuous diffusion language models (DLMs) lag behind discrete counterparts due to the sparse data space and the underexplored design space. In this work, we close this gap with LangFlow, the first continuous DLM to rival discrete diffusion, by connecting embedding-space DLMs to Flow Matching via Bregman divergence, alongside three key innovations: (1) we derive a novel ODE-based NLL bound for principled evaluation of continuous flow-based language models; (2) we propose an information-uniform principle for setting the noise schedule, which motivates a learnable noise scheduler based on a Gumbel distribution; and (3) we revise prior training protocols by incorporating self-conditioning, as we find it improves both likelihood and sample quality of embedding-space DLMs with effects substantially different from discrete diffusion. Putting everything together, LangFlow rivals top discrete DLMs on both the perplexity (PPL) and the generative perplexity (Gen. PPL), reaching a PPL of 30.0 on LM1B and 24.6 on OpenWebText. It even exceeds autoregressive baselines in zero-shot transfer on 4 out of 7 benchmarks. LangFlow provides the first clear evidence that continuous diffusion is a promising paradigm for language modeling. Homepage: https://github.com/nealchen2003/LangFlow
hub tools
citation-role summary
background 1
baseline 1
citation-polarity summary
years
2026 14representative citing papers
Low Gen-PPL in continuous diffusion LMs results from repetition caused by a 1D contractive attractor in self-conditioning feedback; ACE subtracts the direction to reduce repetition to human levels while preserving quality.
Flow models reach 99.2% Sudoku accuracy in 7 passes and 96.1% on out-of-distribution Sudoku-Extreme by selecting dynamically stable candidates and training with self-conditioning plus DPO to avoid failed outputs.
MLFMs combine masking with continuous flows to scale flow-based language models to reasoning and instruction-following tasks on GSM8K and MT-Bench.
Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated as representation-decoder systems.
Marginal-conditioned bridges enable training-free sampling from Flow Language Models by drawing clean one-hot endpoints from factorized posteriors and using Ornstein-Uhlenbeck bridges, preserving token marginals and reducing denoising error versus conditional-mean bridges.
StyleShield uses flow matching in continuous token embeddings with a DiT backbone to achieve 94.6% evasion on trained detectors and over 99% on unseen ones in Chinese benchmarks, with 0.928 semantic similarity, plus a RateAudit method to arbitrarily control detection rates.
FMLM+ with Posterior Refinement bridges masked diffusion and flow map models to match discrete baseline quality in language generation using 32x fewer neural function evaluations via posterior scoring and refinement.
Adapting LLaDA-8B-Instruct via Discrete Stochastic Localization with continuous per-token Gaussian noise yields continuous denoising that achieves top ROUGE-1 on zero-shot summarization at low step budgets and adds selective noisy-state robustness.
DiLaDiff augments masked diffusion LMs with latent space modeling and consistency distillation to improve token correlation capture and inference speed.
RePlaid achieves a 20x compute gap to autoregressive models, new SOTA PPL of 22.1 among continuous DLMs on OpenWebText, and competitive scaling laws by aligning architecture with modern discrete DLMs.
ELF applies continuous-time flow matching in embedding space for language generation and reports outperforming prior discrete and continuous diffusion language models with fewer steps.
Few-step deterministic maps on continuous text latents fail because they cannot resolve discrete branch choices before sharp categorical readouts, with failure governed by decoder sharpness rather than transport accuracy.
High-noise flow-matching training makes one-step VLA action decoding competitive with multi-step decoding because actions are compact targets under rich observations.
citing papers explorer
-
Self-conditioned Flow Map Language Models via Fixed-point Flows
Self-conditioned flow language models solve fixed-point iterations, enabling fixed-point flow maps that distill into FMLM* which outperforms SOTA in few-step generation on OpenWebText.
-
Low Perplexity is Repetition: A One-Dimensional Self-Conditioning Attractor in Continuous Diffusion LMs
Low Gen-PPL in continuous diffusion LMs results from repetition caused by a 1D contractive attractor in self-conditioning feedback; ACE subtracts the direction to reduce repetition to human levels while preserving quality.
-
Flow Reasoning Models: Scaling Reasoning Through Iterative Self-Refinement
Flow models reach 99.2% Sudoku accuracy in 7 passes and 96.1% on out-of-distribution Sudoku-Extreme by selecting dynamically stable candidates and training with self-conditioning plus DPO to avoid failed outputs.
-
Masked Language Flow Models
MLFMs combine masking with continuous flows to scale flow-based language models to reasoning and instruction-following tasks on GSM8K and MT-Bench.
-
Continuous Language Diffusion as a Decoder-Interface Problem
Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated as representation-decoder systems.
-
Sampling from Flow Language Models via Marginal-Conditioned Bridges
Marginal-conditioned bridges enable training-free sampling from Flow Language Models by drawing clean one-hot endpoints from factorized posteriors and using Ornstein-Uhlenbeck bridges, preserving token marginals and reducing denoising error versus conditional-mean bridges.
-
StyleShield: Exposing the Fragility of AIGC Detectors through Continuous Controllable Style Transfer
StyleShield uses flow matching in continuous token embeddings with a DiT backbone to achieve 94.6% evasion on trained detectors and over 99% on unseen ones in Chinese benchmarks, with 0.928 semantic similarity, plus a RateAudit method to arbitrarily control detection rates.
-
Posterior Refinement: Fast Language Generation via Any-Order Flow Maps
FMLM+ with Posterior Refinement bridges masked diffusion and flow map models to match discrete baseline quality in language generation using 32x fewer neural function evaluations via posterior scoring and refinement.
-
DSL-LLaDA: Scaling Continuous Denoising to 8B Masked Diffusion LMs
Adapting LLaDA-8B-Instruct via Discrete Stochastic Localization with continuous per-token Gaussian noise yields continuous denoising that achieves top ROUGE-1 on zero-shot summarization at low step budgets and adds selective noisy-state robustness.
-
DiLaDiff: Distilled Latent-Augmented Diffusion for Language Modeling
DiLaDiff augments masked diffusion LMs with latent space modeling and consistency distillation to improve token correlation capture and inference speed.
-
Continuous Diffusion Scales Competitively with Discrete Diffusion for Language
RePlaid achieves a 20x compute gap to autoregressive models, new SOTA PPL of 22.1 among continuous DLMs on OpenWebText, and competitive scaling laws by aligning architecture with modern discrete DLMs.
-
ELF: Embedded Language Flows
ELF applies continuous-time flow matching in embedding space for language generation and reports outperforming prior discrete and continuous diffusion language models with fewer steps.
-
Why Do Few-Step Text Latents Fail When Image Latents Work? Non-Commitment at Sharp Categorical Readouts
Few-step deterministic maps on continuous text latents fail because they cannot resolve discrete branch choices before sharp categorical readouts, with failure governed by decoder sharpness rather than transport accuracy.
-
Let It Be Simple: One-Step Action Generation for Vision-Language-Action Models
High-noise flow-matching training makes one-step VLA action decoding competitive with multi-step decoding because actions are compact targets under rich observations.