{"kind":"task","effective_mode":"full","benchmark":{"kind":"benchmark","effective_mode":"full","slug":"mgsm","formal_name":"MGSM","introduction":"MGSM is a human translation of 250 GSM8K problems into eleven languages, measuring whether reasoning survives the change of language. The gap against English is the multilingual signal.","introduction_ja":"","introduction_en":"","category":"Category not supplied","task_count":null,"acquisition_status":"Acquisition status not supplied","official_url":"https://github.com/google-research/url-nlp","indexing_mode":"index","profile":{"resources":[],"task_format":"","scoring":"","metric":"","size":"","answer_access":"","license":"","citation":"","maintainer":"","released":"","why_hard":"","related":[],"history":"","construction":"","limitations":"","saturation":"","scores":[]}},"task_id":"5a762e6e-68c2-5783-b906-bf7692c86740","task_key":"th--test--5a762e6e-68c2-5783-b906-bf7692c86740","task_revision_id":"1","upstream_id":"","short_description":"ในรังมีผึ้ง 700 ตัว ผึ้งงานมีจำนวนเป็นสองเท่าของลูกผึ้ง…","config":"th","split":"test","body":"{\"question\":\"ในรังมีผึ้ง 700 ตัว ผึ้งงานมีจำนวนเป็นสองเท่าของลูกผึ้ง และลูกผึ้งมีจำนวนเป็นสองเท่าของนางพญาผึ้ง ในรังมีผึ้งงานกี่ตัว\"}","display_format":"math","language":"","answer_status":"published","assets":[],"source_url":"https://github.com/google-research/url-nlp","history":"initial import","indexing_mode":"index","subproblems":[],"grids":[]}