Papers › Language Models are Few-Shot Learners
Language Models are Few-Shot Learners
Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, Dario Amodei
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.
In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.
Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
For agents, on Syntology's MCP service (how to connect):
get_harvested_code_for_paper(arxiv_id="2005.14165")get_code(code_sha256="…")with a sample'scode_sha256from the list belowget_citation_path(paper_1="2005.14165", paper_2="…")with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="2005.14165")
Code
The archive's official repository (openai/gpt-3) was not harvested; the 65 samples below come from 18 community repositories.
Syntology 65 code samples harvested from 18 repositories linked to this paper: 45 ran (of which 0 constructed an object rather than computing a result; 37 with no instrument failure: 3 honoured, 2 violated, 32 with no contract checked; 8 where Syntology's instrument failed) · 20 unverified. Where Syntology's instrument failed: 7 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.
By repository: community (archive-listed): 65 samples from 18 repositories, 45 ran (37 with no instrument failure; 8 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.
No sample is linked to this paper from 24 of its repositories that Syntology harvested for other papers:
- EleutherAI/lm-evaluation-harness, community (archive-listed): harvested for 2 other papers, listed under the repository
- RUCAIBox/LLMBox, community (archive-listed): harvested for another paper, listed under the repository
- Samyu0304/thought-propagation, community (archive-listed): harvested for another paper, listed under the repository
- allenai/macaw, community (archive-listed): harvested for another paper, listed under the repository
- asahi417/lmppl, community (archive-listed): harvested for another paper, listed under the repository
- bigscience-workshop/promptsource, community (archive-listed): harvested for 2 other papers, listed under the repository
- codedotal/gpt-code-clippy, community (archive-listed): harvested for another paper, listed under the repository
- ethanjperez/true_few_shot, community (archive-listed): harvested for another paper, listed under the repository
- fywalter/label-bias, community (archive-listed): harvested for another paper, listed under the repository
- ggerganov/llama.cpp, community (archive-listed): harvested for another paper, listed under the repository
- ggml-org/llama.cpp, community (archive-listed): harvested for another paper, listed under the repository
- grantslatton/llama.cpp, community (archive-listed): harvested for another paper, listed under the repository
- hazyresearch/ama_prompting, community (archive-listed): harvested for another paper, listed under the repository
- hojjat-mokhtarabadi/promptsource, community (archive-listed): harvested for another paper, listed under the repository
- lambert-x/prolab, community (archive-listed): harvested for another paper, listed under the repository
- longhao-chen/aicas2024, community (archive-listed): harvested for another paper, listed under the repository
- national-center-for-ai-saudi-arabia/lm-evaluation-harness, community (archive-listed): harvested for another paper, listed under the repository
- postech-ami/smile-dataset, community (archive-listed): harvested for another paper, listed under the repository
- scrayish/ML_NLP, community (archive-listed): harvested for another paper, listed under the repository
- smarton-empower/smarton-ai, community (archive-listed): harvested for another paper, listed under the repository
- smile-data/smile, community (archive-listed): harvested for another paper, listed under the repository
- um-arm-lab/efficient-eng-2-ltl, community (archive-listed): harvested for another paper, listed under the repository
- PaddlePaddle/PaddleNLP, community (archive-listed): harvested for another paper, listed under the repository
- mindspore-ai/models, community (archive-listed): harvested for another paper, listed under the repository
Levels written between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 45 samples that ran have no recorded date.
67 repositories listed; official and paper-mentioned ones first.
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.
Code Syntology ran Syntology
65 samples harvested: 45 ran (of which 0 constructed an object rather than computing a result; 37 with no instrument failure: 3 honoured, 2 violated, 32 with no contract checked; 8 where Syntology's instrument failed) · 20 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.
Licence: 7 of the 65 samples are pointer only, meaning Syntology does not serve that copy's text. For 3 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.
Harvested from 18 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. 4 samples whose link to this paper records no repository are listed under one of this paper's own repositories, because identical code was harvested from it for another paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. For 1 sample that ran, the graph records a signature this page does not show: an argument's recorded type or element kind is a printout of an object in Syntology's harness (such as a mock object, a class or a pandas printout), not the type of an input, so it does not describe the call's shape. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.
Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.
Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: 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. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.
A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.
a20ac4af0aec2beb · report
fixture: signature imgs:list/1/Image, chat:list/1/dict, fmt:str/0/str · returned list · level recorded 2026-09-24 72f5b387b64e9d9b · report
fixture: signature attr:str/0/str, args:list/1/str · returned str · level recorded 2026-09-24 00f5e7d9ee6636ba · report
fixture: signature task:MockTask/0/?, limit:float/0/float · returned int · level recorded 2026-09-24 20a7cc804eb22661 · report
fixture: signature attr:str/0/str, args:tuple/1/str · returned str · level recorded, date not recorded 400ea4f64ccce5ef · report
fixture: signature parser:ArgumentParser/0/? · returned Namespace · level recorded 2026-09-22 c44830a7722ff5f7 · report
fixture: signature cache_requests:str/0/str · returned dict · level recorded, date not recorded 2ee40684ab58dcb9 · report
fixture: signature path:str/0/str · returned dict · level recorded 2026-09-24 census: 26 implementations from 30 papers, 11 distinct outputs across 5 buckets, one shared input per bucket relbert/evaluation/analogy_questions.py ran MIT (permissive) ·
3055a2ccdae837c7 · report
fixture: signature a:list/1/float, b:list/1/float, zero_vector_mask:int/0/int · returned float · level recorded 2026-09-24 census: 26 implementations from 30 papers, 11 distinct outputs across 5 buckets, one shared input per bucket relbert/evaluation/relation_mapping.py ran fingerprinted MIT (permissive) ·
3e6ff0c8581ee623 · report
fixture: signature a:ndarray/1/float, b:ndarray/1/float · returned float64 · level recorded 2026-09-24 5307f83d52fe15d1 · report
fixture: signature word_pair:tuple/1/str, template:str/0/str, mask_token:str/0/str · returned str · level recorded 2026-09-24 3866cbd2c9f5fd58 · report
fixture: signature a:list/1/float, b:list/1/float · returned float · level recorded 2026-09-24 fdde66f5d1999801 · report
fixture: signature tensor_positive:Tensor/2/float, tensor_negative:Tensor/2/float, temperature:float/0/float, info_loob:bool/0/bool, linear:callable/0/?, device:str/0/str · returned Tensor · level recorded 2026-09-24 bd67f40255aa71d3 · report
fixture: signature tensor_anchor:Tensor/2/float, tensor_positive:Tensor/2/float, tensor_negative:Tensor/2/float, tensor_positive_parent:Tensor/2/float, tensor_negative_parent:Tensor/2/float, margin:float/0/float, linear:callable/0/?, device:str/0/str · returned Tensor · level recorded 2026-09-24 63b45afd720a808e · report
fixture: signature _list:list/1/int · returned int · level recorded 2026-09-24 99b9d9d7a4d6ca29 · report
fixture: signature original_list:list/1/int, flatten_value:list/1/int · returned list · level recorded 2026-09-24 c7a7a9dca42a1347 · report
fixture: signature _file:str/0/str · returned dict · level recorded 2026-09-24 census: 8 implementations from 8 papers, 3 distinct outputs across 3 buckets, one shared input per bucket relbert/trainer.py ran MIT (permissive) ·
11b1822611370e59 · report
fixture: signature name:str/0/str, data:list/1/list · returned Tensor · level recorded 2026-09-24 34f65890e3e58ec4 · report
87cfebcf19ab803b · report
c7271f2f2ce3c4fa · report
c56b548854fce3e6 · report
8e3012dd13fbb68a · report
5decf8cec7692d54 · report
fixture: signature response:MockResponse/0/?, ctxlen:int/0/int · returned tuple · level recorded 2026-09-24 8cb749113ad26cc3 · report
ecbbe3133a0f68b3 · report
fixture: signature imgs:list/1/Image, chat:list/1/dict, fmt:str/0/str · returned list · level recorded 2026-09-24 0fb487fd2495f33f · report
fixture: signature task:MockTask/0/?, limit:float/0/float · returned int · level recorded 2026-09-24 3e9d38ecacd03fa7 · report
fixture: signature attr:str/0/str, args:list/1/int · returned str · level recorded 2026-09-24 cabc905de3f72170 · report
fixture: signature model:callable/0/?, hook_to_steer:dict/0/? · returned _GeneratorContextManager · level recorded 2026-09-24 cc82794b6ec72130 · report
fixture: signature tensor:Tensor/2/float, multiple:int/0/int, value:int/0/int · returned Tensor · level recorded 2026-09-22 4ff6d4e326d3c0db · report
5036094f3af0dc56 · report
ffcac685bcb61388 · report
fixture: signature imgs:list/1/Image, chat:list/1/dict, fmt:str/0/str · returned list · level recorded 2026-09-25 03edfbaef45e2c28 · report
fixture: signature until:NoneType/0/?, max_gen_toks:int/0/int · returned dict · level recorded 2026-09-25 53354f082d9addc6 · report
fixture: signature logprobs:dict/0/?, context_length:int/0/int · returned tuple · level recorded 2026-09-24 0db35c379a280bd8 · report
fixture: signature task:MockTask/0/?, limit:float/0/float · returned int · level recorded 2026-09-25 26519d22d8af6d48 · report
fixture: signature attr:str/0/str, args:list/1/str · returned str · level recorded 2026-09-25 deba15e38b692b23 · report
fixture: signature lst:list/1/int, sections:int/0/int · returned list · level recorded 2026-09-24 5b315604fb608f3b · report
fixture: signature task:MockTask/0/?, limit:float/0/float · returned int · level recorded 2026-09-24 809bc9208dd99dce · report
fixture: signature task_hierarchy:dict/0/?, results:dict/0/?, tab:int/0/int · returned tuple · level recorded 2026-09-24 451bc0edacfc317d · report
fixture: signature model:callable/0/?, weight_decay:float/0/float, learning_rate:float/0/float, betas:tuple/1/float, eps:float/0/float, state:State/0/? · returned AdamW · level recorded 2026-09-24 b1e0531f8c17d5d6 · report
fixture: signature it:int/0/int, warmup_iters:int/0/int, lr_decay_iters:int/0/int, min_lr:float/0/float, lr:float/0/float · returned float · level recorded 2026-09-24 de43c7cdbf869350 · report
fixture: signature input_string:str/0/str · returned tuple · level recorded 2026-09-24 f6825b4eb5fe78c8 · report
e1d89d7579eaaeec · report
7753000f9482cac5 · report
77347b697708c203 · report
d09226e56c2085ea · report
57e732507bad40db · report
fixture: signature text:str/0/str, sep_char:str/0/str, maxsplit:int/0/int · returned list · level recorded 2026-09-24 89b22f3da55a855c · report
fixture: signature response:dict/0/?, ctxlen:int/0/int · returned tuple · level recorded 2026-09-24 0ee5d4d20cf990bd · report
fixture: signature arg:str/0/str · returned bool · level recorded 2026-09-24 ea06eaae4fc1eaf0 · report
fixture: signature attr:str/0/str, args:tuple/1/int · returned str · level recorded 2026-09-22 7cacc3b113e10007 · report
fixture: returned Mock · level recorded 2026-09-24 dfed6e9291316231 · report
fixture: signature args_string:str/0/str · returned dict · level recorded 2026-09-24 9d752127437e6b8b · report
904fac6af5735d5a · report
fixture: signature module:callable/0/?, log_fn:callable/0/? · returned function · level recorded 2026-09-24 64087979238e1d41 · report
fixture: signature module:callable/0/?, log_fn:callable/0/? · returned function · level recorded 2026-09-24 2e6ad295f92e3d9d · report
fixture: signature number:int/0/int · returned bool · level recorded 2026-09-24 3bb8fcfcb4c12121 · report
e66a197775436632 · report
7ca8b4cd5f558e02 · report
f0fc0ac7590a2d39 · report
fixture: signature input_len:int/0/int, vocab_size:int/0/int, embed_dim:int/0/int, num_heads:int/0/int, num_decoders:int/0/int, fc_dim_factor:int/0/int · returned int · level recorded 2026-09-22 306b51e6dcfb30ef · report
e60f912106425ef4 · report
fixture: signature words:int/0/int, embedding:int/0/int · returned ndarray · level recorded 2026-09-22 Tasks
Results from the paper archive 2025-07-28
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections