A 270M-parameter encoder model (XLM-R) achieves 57.7% macro-F1 on the X-Fact multilingual claim verification benchmark, beating the best tested 7-12B LLM (16.9%) and the prior state of the art (41.9%).
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
When Scale Meets Diversity: Evaluating Language Models on Fine-Grained Multilingual Claim Verification
A 270M-parameter encoder model (XLM-R) achieves 57.7% macro-F1 on the X-Fact multilingual claim verification benchmark, beating the best tested 7-12B LLM (16.9%) and the prior state of the art (41.9%).