Papers › TinyStories: How Small Can Language Models Be and Still Speak Coherent English?

TinyStories: How Small Can Language Models Be and Still Speak Coherent English?

12 May 2023arXiv:2305.07759archive 2025-07-28

Ronen Eldan, Yuanzhi Li

Language models (LMs) are powerful tools for natural language processing, but they often struggle to produce coherent and fluent text when they are small. Models with around 125M parameters such as GPT-Neo (small) or GPT-2 (small) can rarely generate coherent and consistent English text beyond a few words even after extensive training. This raises the question of whether the emergence of the ability to produce coherent English text only occurs at larger scales (with hundreds of millions of parameters or more) and complex architectures (with many layers of global attention). In this work, we introduce TinyStories, a synthetic dataset of short stories that only contain words that a typical 3 to 4-year-olds usually understand, generated by GPT-3.5 and GPT-4. We show that TinyStories can be used to train and evaluate LMs that are much smaller than the state-of-the-art models (below 10 million total parameters), or have much simpler architectures (with only one transformer block), yet still produce fluent and consistent stories with several paragraphs that are diverse and have almost perfect grammar, and demonstrate reasoning capabilities. We also introduce a new paradigm for the evaluation of language models: We suggest a framework which uses GPT-4 to grade the content generated by these models as if those were stories written by students and graded by a (human) teacher. This new paradigm overcomes the flaws of standard benchmarks which often requires the model's output to be very structures, and moreover provides a multidimensional score for the model, providing scores for different capabilities such as grammar, creativity and consistency. We hope that TinyStories can facilitate the development, analysis and research of LMs, especially for low-resource or specialized domains, and shed light on the emergence of language capabilities in LMs.

PaperPDFCodeCode Syntology ran

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):

Code

Syntology 18 code samples harvested from 3 repositories linked to this paper: 10 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 3 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 2 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): 18 samples from 3 repositories, 10 ran (7 with no instrument failure; 3 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 1 of its repositories that Syntology harvested for other papers:

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.

cg80499/kan-gpt-2 mentioned on GitHubjaxSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to KAN: Kolmogorov-Arnold Networks (arXiv:2404.19756; that paper's own run record: 29 ran (of which 9 constructed an object rather than computing a result; 29 with no instrument failure: 1 honoured, 0 violated, 28 with no contract checked; 0 where Syntology's instrument failed) · 16 unverified); nothing checks that they implement this paper's method. report
danbraunai/simple_stories_train mentioned on GitHubjaxMITSyntology: 10 samples, 5 ran (2 instrument-failed) · 5 unverified report
lennart-finke/simple_stories_generate mentioned on GitHubSyntology: not harvested report
noanabeshima/tinymodel mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (1 instrument-failed) · 0 unverified report
phixion/phixion mentioned on GitHubpytorchSyntology: not harvested report
vizuaraai/tiny-stories-regional mentioned on GitHubpytorchMITSyntology: 7 samples, 4 ran (0 instrument-failed) · 3 unverified report
zhaoolee/garss mentioned on GitHubpytorchSyntology: not harvested report
sri9s/tinystories-language-models pytorchSyntology: not harvested report

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

18 samples harvested: 10 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 3 where Syntology's instrument failed) · 8 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.

2ran · our draft was wrong
1ran; the contract check could not run
7ran
8unverified

Licence: 0 of the 18 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 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. “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.

danbraunai/simple_stories_train community (archive-listed) 10 samples, 5 ran (2 instrument-failed) · 5 unverified
convert_dotted_args_to_nested_dict simple_stories_train/utils.py ran MIT (permissive) · b73c48f33832ec9c · report fixture: signature args:dict/0/? · returned dict · level recorded 2026-09-24
create_tokenizer in 8 papers simple_stories_train/tokenizer.py ran MIT (permissive) · ea33e0525958ca7b · report fixture: signature vocab_size:int/0/int · returned Tokenizer · level recorded 2026-09-25
is_checkpoint_step simple_stories_train/utils.py ran fingerprinted MIT (permissive) · fa261019edd7ff74 · report fixture: signature step:int/0/int · returned bool · level recorded 2026-09-24
pad_vocab simple_stories_train/train_gpt2.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant MIT (permissive) · cc82794b6ec72130 · report fixture: signature tensor:Tensor/2/float, multiple:int/0/int, value:int/0/int · returned Tensor · level recorded 2026-09-22
setup_logger in 122 papers simple_stories_train/log.py ran · our draft was wrong MIT (permissive) · d8a05e02085eeb90 · report fixture: returned RootLogger · level recorded 2026-09-24
clean_dataset in 5 papers simple_stories_train/tokenizer.py unverified MIT (permissive) · 69981ccab3e11aa0 · report
create_data_loader in 8 papers simple_stories_train/dataloaders.py unverified MIT (permissive) · 2425cc683fe447c3 · report
load_tokenizer in 66 papers simple_stories_train/tokenizer.py unverified MIT (permissive) · c430eca0b62e5574 · report
replace_pydantic_model simple_stories_train/utils.py unverified MIT (permissive) · a3875bb0a0255e77 · report
tokenize_and_concatenate simple_stories_train/dataloaders.py unverified MIT (permissive) · 2a0bbd4146b3cf66 · report
noanabeshima/tinymodel community (archive-listed) 1 sample, 1 ran (1 instrument-failed) · 0 unverified
parse_mlp_tag tinymodel/lm.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 7f351de9044b032b · report fixture: signature mlp_tag:str/0/str · returned tuple · level recorded 2026-09-22
vizuaraai/tiny-stories-regional community (archive-listed) 7 samples, 4 ran (0 instrument-failed) · 3 unverified
compare_metrics results/compare_models_statistics.py ran MIT (permissive) · 9ca1ae242b588ec1 · report fixture: signature file1_path:str/0/str, file2_path:str/0/str, n_samples:int/0/int, alpha:float/0/float · returned dict · level recorded 2026-09-25
extract_metrics in 5 papers results/compare_models_statistics.py ran MIT (permissive) · e88f4b6f04069085 · report fixture: signature file_path:str/0/str · returned dict · level recorded 2026-09-24
generate_story prompting/requests_helper.py ran MIT (permissive) · 65f2a574a10b1e7b · report fixture: returned str · level recorded 2026-09-24
normalize_text in 60 papers analysis/BERT_BLEU_eval.py ran fingerprinted MIT (permissive) · 31f124b85a98aa6e · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-24
evaluate_story prompting/requests_helper.py unverified MIT (permissive) · f50e36ff8515bb12 · report
get_model_and_tokenizer in 15 papers translation/get_model_token.py unverified MIT (permissive) · 654be48c7a17f9d2 · report
process_chunk in 5 papers prompting/requests_helper.py unverified MIT (permissive) · b062d9bb343787c7 · report

Datasets

Introduced by this paper, per the archive.

TinyStories

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Absolute Position EncodingsAdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDiscriminative Fine-TuningDropoutGPT-2GPT-3GPT-4GPT-NeoLabel SmoothingLayer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformerWeight Decay

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