HiTZ/Llama-3.1-8B-Instruct-multi-truth-judge
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:May 22, 2025License:llama3.1Architecture:Transformer0.0K Cold

HiTZ/Llama-3.1-8B-Instruct-multi-truth-judge is an 8 billion parameter LLM-as-a-Judge model developed by HiTZ Center and affiliated institutions, fine-tuned from Meta-Llama-3.1-8B-Instruct. This model specializes in assessing the truthfulness of text generated by other language models across multiple languages, including English, Basque, Catalan, Galician, and Spanish. It is designed for evaluating model outputs, particularly within the context of multilingual truthfulness benchmarks.

Loading preview...