DiffScore is a bidirectional masked-diffusion evaluation framework that measures text recoverability across masking rates and outperforms autoregressive baselines on ten benchmarks.
ROUGE: A package for automatic evaluation of summaries
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
Case study describing the design and deployment of a RAG system for querying railway engineering regulations.
citing papers explorer
-
DiffScore: Text Evaluation Beyond Autoregressive Likelihood
DiffScore is a bidirectional masked-diffusion evaluation framework that measures text recoverability across masking rates and outperforms autoregressive baselines on ten benchmarks.
-
Retrieval-Augmented Generation to Support Railways Engineering Tasks: A Case Study
Case study describing the design and deployment of a RAG system for querying railway engineering regulations.