{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/unicorn-on-rainbow-a-universal-commonsense","title":"UNICORN on RAINBOW: A Universal Commonsense Reasoning Model on a New Multitask Benchmark","arxiv_id":"2103.13009","date":"2021-03-24","proceeding":null,"authors":["Nicholas Lourie","Ronan Le Bras","Chandra Bhagavatula","Yejin Choi"],"abstract":"Commonsense AI has long been seen as a near impossible goal -- until recently. Now, research interest has sharply increased with an influx of new benchmarks and models. We propose two new ways to evaluate commonsense models, emphasizing their generality on new tasks and building on diverse, recently introduced benchmarks. First, we propose a new multitask benchmark, RAINBOW, to promote research on commonsense models that generalize well over multiple tasks and datasets. Second, we propose a novel evaluation, the cost equivalent curve, that sheds new insight on how the choice of source datasets, pretrained language models, and transfer learning methods impacts performance and data efficiency. We perform extensive experiments -- over 200 experiments encompassing 4800 models -- and report multiple valuable and sometimes surprising findings, e.g., that transfer almost always leads to better or equivalent performance if following a particular recipe, that QA-based commonsense datasets transfer well with each other, while commonsense knowledge graphs do not, and that perhaps counter-intuitively, larger models benefit more from transfer than smaller ones. Last but not least, we introduce a new universal commonsense reasoning model, UNICORN, that establishes new state-of-the-art performance across 8 popular commonsense benchmarks, aNLI (87.3%), CosmosQA (91.8%), HellaSWAG (93.9%), PIQA (90.1%), SocialIQa (83.2%), WinoGrande (86.6%), CycIC (94.0%) and CommonsenseQA (79.3%).","url_abs":"https://arxiv.org/abs/2103.13009v1","url_pdf":"https://arxiv.org/pdf/2103.13009v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"unicorn-on-rainbow-a-universal-commonsense","repo_url":"https://github.com/allenai/rainbow","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"common-sense-reasoning","task_name":"Common Sense Reasoning"},{"task_slug":"hellaswag","task_name":"HellaSwag"},{"task_slug":"knowledge-graphs","task_name":"Knowledge Graphs"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"sentence-completion","task_name":"Sentence Completion"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"winogrande","task_name":"Winogrande"}],"methods":[],"datasets_introduced":[{"slug":"rainbow","name":"Rainbow","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/common-sense-reasoning-on-commonsenseqa","task":"Common Sense Reasoning","dataset":"CommonsenseQA","model":"Unicorn 11B (fine-tuned)","rank_in_archive_order":6,"of":38,"metrics":{"Accuracy":"79.3"},"uses_additional_data":false},{"leaderboard":"/sota/common-sense-reasoning-on-winogrande","task":"Common Sense Reasoning","dataset":"WinoGrande","model":"Unicorn 11B (fine-tuned)","rank_in_archive_order":2,"of":77,"metrics":{"Accuracy":"91.3"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-piqa","task":"Question Answering","dataset":"PIQA","model":"Unicorn 11B (fine-tuned)","rank_in_archive_order":1,"of":67,"metrics":{"Accuracy":"90.1"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-social-iqa","task":"Question Answering","dataset":"SIQA","model":"Unicorn 11B (fine-tuned)","rank_in_archive_order":1,"of":24,"metrics":{"Accuracy":"83.2"},"uses_additional_data":false},{"leaderboard":"/sota/sentence-completion-on-hellaswag","task":"Sentence Completion","dataset":"HellaSwag","model":"Unicorn 11B (fine-tuned)","rank_in_archive_order":8,"of":89,"metrics":{"Accuracy":"93.9"},"uses_additional_data":true}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2103.13009","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2103.13009"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/allenai/rainbow","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"unverified":5},"by_repo_kind":{"official":{"samples":5,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"40c05820ad047307","entry":"configure_logging","repo":"allenai/rainbow","repo_kind":"official","path":"src/rainbow/utils.py","file_url":"https://github.com/allenai/rainbow/blob/HEAD/src/rainbow/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"40c05820ad047307"}},{"code_sha256_prefix":"86dcf04053de697c","entry":"data","repo":"allenai/rainbow","repo_kind":"official","path":"src/rainbow/utils.py","file_url":"https://github.com/allenai/rainbow/blob/HEAD/src/rainbow/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"86dcf04053de697c"}},{"code_sha256_prefix":"21730ebd53e72357","entry":"make_add_field_names_preprocessor","repo":"allenai/rainbow","repo_kind":"official","path":"src/rainbow/preprocessors.py","file_url":"https://github.com/allenai/rainbow/blob/HEAD/src/rainbow/preprocessors.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"21730ebd53e72357"}},{"code_sha256_prefix":"0d51b04143d199db","entry":"make_filter_preprocessor","repo":"allenai/rainbow","repo_kind":"official","path":"src/rainbow/preprocessors.py","file_url":"https://github.com/allenai/rainbow/blob/HEAD/src/rainbow/preprocessors.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0d51b04143d199db"}},{"code_sha256_prefix":"bba4a76bdd42388e","entry":"string_to_seed","repo":"allenai/rainbow","repo_kind":"official","path":"src/rainbow/utils.py","file_url":"https://github.com/allenai/rainbow/blob/HEAD/src/rainbow/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"bba4a76bdd42388e"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}