Papers › SelfCodeAlign: Self-Alignment for Code Generation

SelfCodeAlign: Self-Alignment for Code Generation

31 Oct 2024arXiv:2410.24198archive 2025-07-28

Yuxiang Wei, Federico Cassano, Jiawei Liu, Yifeng Ding, Naman jain, Zachary Mueller, Harm de Vries, Leandro von Werra, Arjun Guha, Lingming Zhang

Instruction tuning is a supervised fine-tuning approach that significantly improves the ability of large language models (LLMs) to follow human instructions. We propose SelfCodeAlign, the first fully transparent and permissive pipeline for self-aligning code LLMs without extensive human annotations or distillation. SelfCodeAlign employs the same base model for inference throughout the data generation process. It first extracts diverse coding concepts from high-quality seed snippets to generate new tasks. It then samples multiple responses per task, pairs each with test cases, and validates them in a sandbox environment. Finally, passing examples are selected for instruction tuning. In our primary experiments, we use SelfCodeAlign with CodeQwen1.5-7B to generate a dataset of 74k instruction-response pairs. Finetuning on this dataset leads to a model that achieves a 67.1 pass@1 on HumanEval+, surpassing CodeLlama-70B-Instruct despite being ten times smaller. Across all benchmarks, this finetuned model consistently outperforms the original version trained with OctoPack, the previous state-of-the-art method for instruction tuning without human annotations or distillation. Additionally, we show that SelfCodeAlign is effective across LLMs of various sizes, from 3B to 33B, and that the base models can benefit more from alignment with their own data distribution. We further validate each component's effectiveness in our pipeline, showing that SelfCodeAlign outperforms both direct distillation from GPT-4o and leading GPT-3.5-based distillation methods, such as OSS-Instruct and Evol-Instruct. SelfCodeAlign has also led to the creation of StarCoder2-Instruct, the first fully transparent, permissively licensed, and self-aligned code LLM that achieves state-of-the-art coding performance.

PaperPDFCodeCode Syntology ranMethod as code

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 37 code samples harvested from 3 repositories linked to this paper: 30 ran (of which 3 constructed an object rather than computing a result; 21 with no instrument failure: 1 honoured, 0 violated, 20 with no contract checked; 9 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 9 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 12 samples from 1 repository, 9 ran (3 with no instrument failure; 6 where Syntology's instrument failed); community (archive-listed): 22 samples from 1 repository, 18 ran (17 with no instrument failure; 1 where Syntology's instrument failed); community: 3 samples from 1 repository, 3 ran (1 with no instrument failure; 2 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.

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 5 of the 30 samples that ran have no recorded date.

bigcode-project/selfcodealign officialmentioned in papermentioned on GitHubSyntology: 12 samples, 9 ran (6 instrument-failed) · 3 unverified report
bigcode-project/starcoder2-self-align mentioned on GitHubApache-2.0Syntology: 22 samples, 18 ran (1 instrument-failed) · 4 unverified 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.

Code Syntology ran Syntology

37 samples harvested: 30 ran (of which 3 constructed an object rather than computing a result; 21 with no instrument failure: 1 honoured, 0 violated, 20 with no contract checked; 9 where Syntology's instrument failed) · 7 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.

1ran · honoured contract
9ran · our draft was wrong
20ran
7unverified

Licence: 0 of the 37 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”. “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.

bigcode-project/selfcodealign official repository 12 samples, 9 ran (6 instrument-failed) · 3 unverified
Args in 6 papers src/star_align/self_ossinstruct.py ran Apache-2.0 (permissive) · ae0b66e2959a4a8b · report fixture: returned Args · level recorded 2026-09-22
Example src/star_align/self_ossinstruct.py ran Apache-2.0 (permissive) · ce4e9324e38a0ddf · report fixture: returned Example · level recorded 2026-09-22
Property src/star_align/self_ossinstruct.py ran Apache-2.0 (permissive) · 5b83b230eb85881a · report fixture: returned Property · level recorded 2026-09-22
build_kwargs src/star_align/self_ossinstruct.py ran · our draft was wrong Apache-2.0 (permissive) · 8dd50472f41a7528 · report fixture: signature instruct_mode:str/0/str, example:dict/0/? · returned dict · level recorded 2026-09-22
flatten_openai_responses src/star_align/self_ossinstruct.py ran · our draft was wrong Apache-2.0 (permissive) · 1fc3a1bcc54c9a56 · report fixture: signature responses:list/1/Completion · returned list · level recorded 2026-09-22
parse_generated_content src/star_align/self_ossinstruct.py ran · our draft was wrong Apache-2.0 (permissive) · a75703b3e4e8a6a7 · report fixture: signature content:str/0/str, instruct_mode:str/0/str · returned dict · level recorded 2026-09-22
parse_property src/star_align/self_ossinstruct.py ran · our draft was wrong Apache-2.0 (permissive) · fba12f4a4a9f3c28 · report fixture: signature content:str/0/str · returned Property · level recorded 2026-09-22
postprocess in 70 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 2 buckets, one shared input per bucket
evaluation/ds_1000.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 1173d57fe1b4d158 · report fixture: signature text:str/0/str · returned str · level recorded, date not recorded
preprocess_completion_prompt evaluation/ds_1000.py ran · our draft was wrong Apache-2.0 (permissive) · 956cd0aa1cec0821 · report fixture: signature prompt:str/0/str · returned tuple · level recorded, date not recorded
Fewshot src/star_align/self_ossinstruct.py unverified Apache-2.0 (permissive) · 69d27e542d70e540 · report
get_ossinstruct_fewshots src/star_align/self_ossinstruct.py unverified Apache-2.0 (permissive) · d063f1765c316bac · report
main in 235 papers src/star_align/self_ossinstruct.py unverified Apache-2.0 (permissive) · 5c02c0de38a6ae98 · report
bigcode-project/starcoder2-self-align community (archive-listed) 22 samples, 18 ran (1 instrument-failed) · 4 unverified
benchmark_name_to_filter_reason src/star_align/decontamination/find_substrings.py ran fingerprinted Apache-2.0 (permissive) · a349a736c168650a · report fixture: signature benchmark_name:str/0/str · returned str · level recorded 2026-09-24
chunked in 5 papers src/star_align/utils.py ran Apache-2.0 (permissive) · a6974ed3c2ddb7eb · report fixture: signature seq:list/1/int, n:int/0/int · returned generator · level recorded 2026-09-24
extract_and_concat_function_names src/star_align/clean_data.py ran fingerprinted Apache-2.0 (permissive) · ab5fa2503ae7af82 · report fixture: signature python_content:str/0/str · returned str · level recorded 2026-09-24
extract_code_with_indentation src/star_align/collect_snippets.py ran Apache-2.0 (permissive) · 5483cff90bd19eb8 · report fixture: signature node:MockNode/0/?, source_code:str/0/str · returned str · level recorded 2026-09-24
extract_docstring in 12 papers src/star_align/decontamination/benchmark_data.py ran · our draft was wrong Apache-2.0 (permissive) · 0edeacd7a9034f53 · report fixture: signature prompt:str/0/str · returned str · level recorded 2026-09-24
extract_ds_1000_prompt src/star_align/decontamination/benchmark_data.py ran Apache-2.0 (permissive) · 2160a71414dde9a6 · report fixture: signature prompt:str/0/str · returned str · level recorded 2026-09-25
filter_reason_to_benchmark_name src/star_align/decontamination/find_substrings.py ran Apache-2.0 (permissive) · 9c5b01ef26d74c74 · report fixture: signature filter_reason:str/0/str · returned str · level recorded 2026-09-25
form_starcoder_infill src/star_align/llm_wrapper.py ran fingerprinted Apache-2.0 (permissive) · 971c70adc27358da · report fixture: signature prefix:str/0/str, suffix:str/0/str · returned str · level recorded 2026-09-26
make_python_membound_code_prefix src/star_align/execution_filter.py ran fingerprinted Apache-2.0 (permissive) · 8c087c3185ff5fab · report fixture: signature limit_mb:int/0/int · returned str · level recorded 2026-09-24
ngrams in 18 papers src/star_align/minhash_dedup.py ran Apache-2.0 (permissive) · a0d7edeceb109990 · report fixture: signature sequence:list/1/str, n:int/0/int, min_length:int/0/int · returned zip · level recorded 2026-09-24
pad_sequences in 19 papers src/star_align/llm_wrapper.py ran Apache-2.0 (permissive) · 7b178c461482e3f6 · report fixture: signature sequences:list/1/list, pad_value:int/0/int, padding_side:str/0/str, dtype:dtype/0/?, padding_length:NoneType/0/? · returned Tensor · level recorded 2026-09-26
read_jsonl in 136 papers src/star_align/utils.py ran Apache-2.0 (permissive) · b0fdb2b203090d2e · report fixture: signature path:PosixPath/0/? · returned list · level recorded 2026-09-24
retry in 14 papers src/star_align/utils.py ran Apache-2.0 (permissive) · e10de9c97d0ec5c5 · report fixture: signature errors:tuple/1/callable, max_attempts:int/0/int · returned function · level recorded 2026-09-24
sha1_hash src/star_align/minhash_dedup.py ran Apache-2.0 (permissive) · e5358426087cdd76 · report fixture: signature data:bytes/0/?, d:int/0/int · returned int · level recorded 2026-09-24
split_llama3_response_tests src/star_align/sanitize_data.py ran fingerprinted Apache-2.0 (permissive) · 13b161a5bbc15829 · report fixture: signature response:str/0/str · returned list · level recorded 2026-09-24
suppress_output src/star_align/execution_filter.py ran Apache-2.0 (permissive) · d2c4d113d519cc11 · report fixture: signature func:callable/0/? · returned function · level recorded 2026-09-25
update_benchmark_dict src/star_align/decontamination/find_substrings.py ran Apache-2.0 (permissive) · 303bc7228b4758e6 · report fixture: signature filter_out:dict/0/?, benchmark_cache:str/0/str, excluded_data_cache:str/0/str · returned tuple · level recorded 2026-09-24
xxh3_16hash src/star_align/minhash_dedup.py ran Apache-2.0 (permissive) · b85036e1fe10ab3b · report fixture: signature data:bytes/0/?, seed:int/0/int · returned int · level recorded 2026-09-24
containerized_run src/star_align/execution_filter.py unverified Apache-2.0 (permissive) · 17b4339501de907f · report
extract_snippets_with_constraints src/star_align/collect_snippets.py unverified Apache-2.0 (permissive) · 465ba803e1fe81b9 · report
get_model_context src/star_align/llm_wrapper.py unverified Apache-2.0 (permissive) · 1415e89bd7265711 · report
load_dataset_column in 10 papers src/star_align/decontamination/benchmark_data.py unverified Apache-2.0 (permissive) · 9c7f0c27b8d08b0f · report
sahil280114/codealpaca community 3 samples, 3 ran (2 instrument-failed) · 0 unverified
encode_prompt in 54 papers generate_instruction.py ran · our draft was wrong Apache-2.0 (permissive) · 9b2dd0265fc6e6b6 · report fixture: signature prompt_instructions:list/1/dict · returned str · level recorded, date not recorded
find_word_in_string in 12 papers generate_instruction.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 8ca6a5f9f77053c2 · report fixture: signature w:str/0/str, s:str/0/str · returned Match · level recorded, date not recorded
post_process_gpt3_response in 12 papers generate_instruction.py ran · honoured contract Apache-2.0 (permissive) · a5e4ad4e11cf2143 · report fixture: signature num_prompt_instructions:int/0/int, response:dict/0/? · returned list · level recorded, date not recorded

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for SelfCodeAlign (graph method: SelfCodeAlign). 10 bricks; individually on a fixture: 7 ran (of which 3 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 4 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

  1. main entry · level 0 · unverified · 5c02c0de38a6ae98
  2. Args data_transform · level 1 · ran · ae0b66e2959a4a8b
  3. build_kwargs helper · level 1 · ran · our draft was wrong · 8dd50472f41a7528
  4. flatten_openai_responses helper · level 1 · ran · our draft was wrong · 1fc3a1bcc54c9a56
  5. get_ossinstruct_fewshots helper · level 0 · unverified · d063f1765c316bac
  6. parse_generated_content helper · level 1 · ran · our draft was wrong · a75703b3e4e8a6a7
  7. Property data_transform · level 1 · ran · 5b83b230eb85881a
  8. Example data_transform · level 1 · ran · ce4e9324e38a0ddf
  9. Fewshot data_transform · level 0 · unverified · 69d27e542d70e540
  10. parse_property helper · level 1 · ran · our draft was wrong · fba12f4a4a9f3c28

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2410.24198") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Code GenerationHumanEval

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBASEBPECosine AnnealingDense ConnectionsDropoutGPT-3Layer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionResidual ConnectionSoftmaxWeight 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