{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/improving-language-understanding-by","title":"Improving Language Understanding by Generative Pre-Training","arxiv_id":null,"date":"2018-06-11","proceeding":"Preprint 2018 6","authors":["Alec Radford","Karthik Narasimhan","Tim Salimans","Ilya Sutskever"],"abstract":"Natural language understanding comprises a wide range of diverse tasks such\r\nas textual entailment, question answering, semantic similarity assessment, and\r\ndocument classification. Although large unlabeled text corpora are abundant,\r\nlabeled data for learning these specific tasks is scarce, making it challenging for\r\ndiscriminatively trained models to perform adequately. We demonstrate that large\r\ngains on these tasks can be realized by generative pre-training of a language model\r\non a diverse corpus of unlabeled text, followed by discriminative fine-tuning on each\r\nspecific task. In contrast to previous approaches, we make use of task-aware input\r\ntransformations during fine-tuning to achieve effective transfer while requiring\r\nminimal changes to the model architecture. We demonstrate the effectiveness of\r\nour approach on a wide range of benchmarks for natural language understanding.\r\nOur general task-agnostic model outperforms discriminatively trained models that\r\nuse architectures specifically crafted for each task, significantly improving upon the\r\nstate of the art in 9 out of the 12 tasks studied. For instance, we achieve absolute\r\nimprovements of 8.9% on commonsense reasoning (Stories Cloze Test), 5.7% on\r\nquestion answering (RACE), and 1.5% on textual entailment (MultiNLI).","url_abs":"https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf","url_pdf":"https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.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":"improving-language-understanding-by","repo_url":"https://github.com/2023-MindSpore-1/ms-code-152","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/Mind23-2/MindCode-139","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/MindSpore-scientific/code-7/tree/main/gpt_j","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/PaddlePaddle/FleetX/tree/develop/examples/static/gpt","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"paddle","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/abhaskumarsinha/Corpus2GPT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/abhaskumarsinha/Keras-implementation-of-Transformer-Architecture","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/abhaskumarsinha/MinimalGPT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/huggingface/transformers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/lattice-ai/GPT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/lvyufeng/bert4ms","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/milmor/GPT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/openai/finetune-transformer-lm","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"improving-language-understanding-by","repo_url":"https://github.com/xiuyu0000/new_papers_codes/tree/main/gpt","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"cloze-test","task_name":"Cloze Test"},{"task_slug":"document-classification","task_name":"Document Classification"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"natural-language-inference","task_name":"Natural Language Inference"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"semantic-similarity","task_name":"Semantic Similarity"},{"task_slug":"semantic-textual-similarity","task_name":"Semantic Textual Similarity"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"cosine-annealing","method_name":"Cosine Annealing"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"discriminative-fine-tuning","method_name":"Discriminative Fine-Tuning"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt","method_name":"GPT"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-cosine-annealing","method_name":"Linear Warmup With Cosine Annealing"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[{"slug":"gpt","name":"GPT","full_name":"GPT"}],"results":[{"leaderboard":"/sota/natural-language-inference-on-multinli","task":"Natural Language Inference","dataset":"MultiNLI","model":"Finetuned Transformer LM","rank_in_archive_order":41,"of":67,"metrics":{"Matched":"82.1","Mismatched":"81.4"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-inference-on-snli","task":"Natural Language Inference","dataset":"SNLI","model":"Fine-Tuned LM-Pretrained Transformer","rank_in_archive_order":15,"of":98,"metrics":{"% Test Accuracy":"89.9","% Train Accuracy":"96.6","Parameters":"85m"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-inference-on-scitail","task":"Natural Language Inference","dataset":"SciTail","model":"Finetuned Transformer LM","rank_in_archive_order":4,"of":13,"metrics":{"Accuracy":"88.3"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-race","task":"Question Answering","dataset":"RACE","model":"Finetuned Transformer LM","rank_in_archive_order":4,"of":7,"metrics":{"RACE":"59.0","RACE-h":"57.4","RACE-m":"62.9"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-storycloze","task":"Question Answering","dataset":"StoryCloze","model":"Finetuned Transformer LM","rank_in_archive_order":8,"of":23,"metrics":{"Accuracy":"86.5"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers","held":null,"reason":"no_arxiv_id","checked_against":null,"by_repo":[{"repo":"2023-MindSpore-1/ms-code-152","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"Mind23-2/MindCode-139","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"MindSpore-scientific/code-7","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"PaddlePaddle/FleetX","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2205.10034"],"harvested_for_calls":[{"arxiv_id":"2205.10034","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2205.10034"},"call":"get_harvested_code_for_paper(arxiv_id=\"2205.10034\")"}]},{"repo":"abhaskumarsinha/Corpus2GPT","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":2,"harvested_for":["2005.14165","2302.13971"],"harvested_for_calls":[{"arxiv_id":"2005.14165","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2005.14165"},"call":"get_harvested_code_for_paper(arxiv_id=\"2005.14165\")"},{"arxiv_id":"2302.13971","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2302.13971"},"call":"get_harvested_code_for_paper(arxiv_id=\"2302.13971\")"}]},{"repo":"abhaskumarsinha/Keras-implementation-of-Transformer-Architecture","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["1706.03762"],"harvested_for_calls":[{"arxiv_id":"1706.03762","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1706.03762"},"call":"get_harvested_code_for_paper(arxiv_id=\"1706.03762\")"}]},{"repo":"abhaskumarsinha/MinimalGPT","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":2,"harvested_for":["1706.03762","2005.14165"],"harvested_for_calls":[{"arxiv_id":"1706.03762","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1706.03762"},"call":"get_harvested_code_for_paper(arxiv_id=\"1706.03762\")"},{"arxiv_id":"2005.14165","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2005.14165"},"call":"get_harvested_code_for_paper(arxiv_id=\"2005.14165\")"}]},{"repo":"huggingface/transformers","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":4,"harvested_for":["1706.03762","1910.03771","2006.03654","2010.01057"],"harvested_for_calls":[{"arxiv_id":"1706.03762","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1706.03762"},"call":"get_harvested_code_for_paper(arxiv_id=\"1706.03762\")"},{"arxiv_id":"1910.03771","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1910.03771"},"call":"get_harvested_code_for_paper(arxiv_id=\"1910.03771\")"},{"arxiv_id":"2006.03654","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2006.03654"},"call":"get_harvested_code_for_paper(arxiv_id=\"2006.03654\")"},{"arxiv_id":"2010.01057","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2010.01057"},"call":"get_harvested_code_for_paper(arxiv_id=\"2010.01057\")"}]},{"repo":"lattice-ai/GPT","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"lvyufeng/bert4ms","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["1907.11692"],"harvested_for_calls":[{"arxiv_id":"1907.11692","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1907.11692"},"call":"get_harvested_code_for_paper(arxiv_id=\"1907.11692\")"}]},{"repo":"milmor/GPT","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"openai/finetune-transformer-lm","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"xiuyu0000/new_papers_codes","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_harvested_for_is":"harvested_for: the arXiv ids (as Syntology's graph holds them) of every paper whose samples Syntology's graph links to this repository, in arXiv-id order, complete; the page names the first 3 and says how many more. A link only: it does not say the papers are the same work, that one copied the other, or that the code implements this paper's method; harvested_for_calls gives the get_harvested_code_for_paper call for each arXiv-shaped id among them","by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"924ee7a114d0a1eeca792888415108f5357d725c6945292e5f69315bcc3040e5","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}