Papers › LEVER: Learning to Verify Language-to-Code Generation with Execution

LEVER: Learning to Verify Language-to-Code Generation with Execution

16 Feb 2023arXiv:2302.08468archive 2025-07-28

Ansong Ni, Srini Iyer, Dragomir Radev, Ves Stoyanov, Wen-tau Yih, Sida I. Wang, Xi Victoria Lin

The advent of large language models trained on code (code LLMs) has led to significant progress in language-to-code generation. State-of-the-art approaches in this area combine LLM decoding with sample pruning and reranking using test cases or heuristics based on the execution results. However, it is challenging to obtain test cases for many real-world language-to-code applications, and heuristics cannot well capture the semantic features of the execution results, such as data type and value range, which often indicates the correctness of the program. In this work, we propose LEVER, a simple approach to improve language-to-code generation by learning to verify the generated programs with their execution results. Specifically, we train verifiers to determine whether a program sampled from the LLMs is correct or not based on the natural language input, the program itself and its execution results. The sampled programs are reranked by combining the verification score with the LLM generation probability, and marginalizing over programs with the same execution results. On four datasets across the domains of table QA, math QA and basic Python programming, LEVER consistently improves over the base code LLMs(4.6% to 10.9% with code-davinci-002) and achieves new state-of-the-art results on all of them.

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 22 code samples harvested from 1 repository linked to this paper: 18 ran (of which 0 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 2 violated, 12 with no contract checked; 4 where Syntology's instrument failed) · 4 unverified. Where Syntology's instrument failed: 3 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: official repository: 22 samples from 1 repository, 18 ran (14 with no instrument failure; 4 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 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

niansong1996/lever officialmentioned in papermentioned on GitHubpytorchMITSyntology: 22 samples, 18 ran (4 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

22 samples harvested: 18 ran (of which 0 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 2 violated, 12 with no contract checked; 4 where Syntology's instrument failed) · 4 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 · violated contract
3ran · our draft was wrong
1ran; the contract check could not run
12ran
4unverified

Licence: 1 of the 22 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, 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 niansong1996/lever. “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.

niansong1996/lever official repository 22 samples, 18 ran (4 instrument-failed) · 4 unverified
byte_idx_to_char_idx execution/program_tracing.py ran fingerprinted MIT (permissive) · 7e87fa354c8a130c · report fixture: signature byte_idx:int/0/int, line:str/0/str · returned int · level recorded 2026-09-25
canonicalize_var_dict execution/safe_execution_util.py ran MIT (permissive) · 5601a3dc0fc200e8 · report fixture: signature var_dict:dict/0/? · returned dict · level recorded 2026-09-25
direct_categorize execution/spider_execution.py ran MIT (permissive) · 51262b43d9045912 · report fixture: signature example:dict/0/? · returned list · level recorded 2026-09-25
estimate_pass_at_k in 28 papers few_shot/codex.py ran fingerprinted MIT (permissive) · 3a39b10c8a429dc8 · report fixture: signature n:int/0/int, c:int/0/int, k:int/0/int · returned float64 · level recorded 2026-09-25
is_number in 70 papers execution/wtq_eval.py ran · violated contract MIT (permissive) here; the stored copy is pointer only · e18607dae38b30e2 · report fixture: signature s:str/0/str · returned bool · level recorded 2026-09-22
join_tokens in 7 papers execution/parse.py ran · our draft was wrong MIT (permissive) · 2178b95c919ae420 · report fixture: signature tokens:list/1/Token · returned str · level recorded 2026-09-24
normalize in 638 papers
census: 70 implementations from 82 papers, 51 distinct outputs across 11 buckets, one shared input per bucket
execution/wtq_eval.py ran fingerprinted MIT (permissive) · 5d30efa67a78bf91 · report fixture: signature x:str/0/str · returned str · level recorded 2026-09-25
openai_call few_shot/codex.py ran MIT (permissive) · 73683dd2b9714a13 · report fixture: signature prompts:list/1/str, engine:str/0/str, max_tokens:int/0/int, temperature:float/0/float, top_p:float/0/float, n:int/0/int, best_of:int/0/int, wait_on_limit:bool/0/bool, stop:str/0/str, get_raw_generation_results:bool/0/bool · returned list · level recorded 2026-09-25
pd_df_to_dict execution/spider_execution.py ran MIT (permissive) · 96389cb2ab61121a · report fixture: signature df:DataFrame/0/? · returned tuple · level recorded 2026-09-25
permute_tuple in 7 papers execution/spider_official_exec_match.py ran; the contract check could not run (RAISES) MIT (permissive) · f23d6ff3772963bc · report fixture: signature element:tuple/1/int, perm:tuple/1/int · returned tuple · level recorded 2026-09-24
postprocess in 70 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 2 buckets, one shared input per bucket
execution/parse.py ran · our draft was wrong fingerprinted MIT (permissive) · 58e472cff9780d63 · report fixture: signature query:str/0/str · returned str · level recorded 2026-09-24
python_program_len execution/executors.py ran fingerprinted MIT (permissive) · 48ddf72e21e4525b · report fixture: signature code:str/0/str · returned int · level recorded 2026-09-25
quick_rej in 7 papers execution/spider_official_exec_match.py ran · violated contract MIT (permissive) · a40510ba6a37f798 · report fixture: signature result1:list/1/tuple, result2:list/1/tuple, order_matters:bool/0/bool · returned bool · level recorded 2026-09-24
saved_promptify_mathqa finetuning/lightning_modules/datasets/mathqa_reader.py ran MIT (permissive) · acdd63b83770726f · report fixture: signature prompt_file_path:str/0/str, example:dict/0/?, max_prompt_examples:int/0/int · returned str · level recorded 2026-09-25
simple_canonicalize_var_dict execution/safe_execution_util.py ran MIT (permissive) · f7dbd2bcc2e00ee0 · report fixture: signature var_dict:dict/0/? · returned dict · level recorded 2026-09-25
sql_program_len execution/executors.py ran fingerprinted MIT (permissive) · 53b15e547590af1b · report fixture: signature code:str/0/str · returned int · level recorded 2026-09-25
state_answer_only_func finetuning/lightning_modules/datasets/mathqa_reader.py ran MIT (permissive) · 4269741e408c7847 · report fixture: signature state_dict:dict/0/? · returned str · level recorded 2026-09-25
unorder_row in 7 papers execution/spider_official_exec_match.py ran · our draft was wrong MIT (permissive) · afc8ef89576ef2aa · report fixture: signature row:tuple/1/str · returned tuple · level recorded 2026-09-24
execute in 12 papers execution/safe_execution_util.py unverified MIT (permissive) · 0e6bb17163d0c70b · report
pd_df_from_dict execution/spider_execution.py unverified MIT (permissive) · d4c44c1affe7cfe5 · report
state_simple_str_func finetuning/lightning_modules/datasets/mathqa_reader.py unverified MIT (permissive) · bc02ec961672d267 · report
to_value execution/wtq_eval.py unverified MIT (permissive) · efd6f533b8606dca · report

Tasks

Arithmetic ReasoningCode GenerationMathRerankingSemantic ParsingText-To-SQL

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Arithmetic Reasoning GSM8K code-davinci-002 175B (LEVER, 8-shot) Accuracy 84.5 #50 of 164 Archive leaderboard report
Arithmetic Reasoning GSM8K code-davinci-002 175B (LEVER, 8-shot) Parameters (Billion) 175 #50 of 164 Archive leaderboard report
Code Generation MBPP code-davinci-002 175B + LEVER Accuracy 68.9 #33 of 99 Archive leaderboard report
Semantic Parsing WikiTableQuestions LEVER Accuracy (Dev) 64.6 #11 of 22 Archive leaderboard report
Semantic Parsing WikiTableQuestions LEVER Accuracy (Test) 65.8 #11 of 22 Archive leaderboard report
Semantic Parsing spider code-davinci-002 175B (LEVER) Accuracy 81.9 #2 of 10 Archive leaderboard report
Text-To-SQL spider code-davinci-002 175B (LEVER) Execution Accuracy (Dev) 81.9 #17 of 20 Archive leaderboard report

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

BASEPruningTest

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