A Diagnostic Study of Faithfulness in Diffusion Language Models for Table-to-Text Generation
Diffusion language models generate by iteratively denoising a masked sequence under bidirectional attention, without the strict left-to-right commitment of autoregressive decoding.
ProofPaper ↗
Key points
- Whether this difference affects how reliably they remain grounded in structured inputs is largely unexplored.
- We study this question in table-to-text generation using ToTTo, where highlighted source cells make factual grounding directly testable.
- We compare the masked diffusion models MDLM and LLaDA with GPT-2 and GPT-Neo under an identical retrieval-based few-shot protocol, together with a QLoRA-adapted LLaDA, and we score reference similarity and source grounding as separate axes.
- Reference-based metrics alone are insufficient to assess faithfulness in table-to-text generation.
Sources (1)
- [1]A Diagnostic Study of Faithfulness in Diffusion Language Models for Table-to-Text GenerationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 11:42 AM
Diffusion language models generate by iteratively denoising a masked sequence under bidirectional attention, without the strict left-to-right commitment of autoregressive decoding.
Whether this difference affects how reliably they remain grounded in structured inputs is largely unexplored.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 21, 2026Looped Uniform Diffusion Language Models
- Oct 20, 2026Voxtral 3
- Oct 20, 2026DiffResearch: Accelerating Deep Research with Diffusion Language Models
- Oct 20, 2026Large Flow Language Models
- Oct 19, 2026Amazon SageMaker JumpStart: 2026 year-to-date in review
- Oct 19, 2026BERT Is Already a Denoiser: Cheaply Unlocking Iterative Generation in Masked Language Models
