MultiSocial
we propose the first multilingual (22 languages) and multi-platform (5 social media platforms) dataset for benchmarking machine-generated text detection in the social-media domain, called MultiSocial.
Can I use it?
No license is specified. Ask the authors before using it. Request access on the resource’s page.
About this resource
we propose the first multilingual (22 languages) and multi-platform (5 social media platforms) dataset for benchmarking machine-generated text detection in the social-media domain, called MultiSocial. It contains 472,097 texts, of which about 58k are human-written and approximately the same amount is generated by each of 7 multilingual LLMs - aya-101, Mistral-7B-Instruct-v0.2, gpt-3.5-turbo-0125, v5-Eagle-7B-HF, vicuna-13b, opt-iml-max-30b, Llama-2-70b-chat-hf
More resources for this task
All resources for “Check text and models” (3)- DatasetGEST
GEST – a manually created dataset designed to measure gender-stereotypical reasoning in language models and machine translation systems.
- NLP KInIT
- WUDAP KInIT
- Ethics KInIT
Apache 2.0 - DatasetChiby
The CHIBY error corpus is based on Slovak Wikipedia revisions and features paragraph-level text edits with automatically annotated change types designed to maximize data…
No license specified