Pith. sign in

REVIEW 5 major objections 5 minor 2 cited by

Automating Financial Statement Audits with Large Language Models

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read GPT-4 flags financial statement errors perfectly but completes a full audit only 4.1% of the time.

desk verdict A useful five-stage evaluation framework for LLM auditing, but the perfect detection scores may be an artifact of row-aligned synthetic data, and the evaluation lacks baselines and significance testing. read the letter →

arxiv 2506.17282 v1 pith:ILI7MAI5 submitted 2025-06-14 cs.IR

Rushi Wang , Jiateng Liu , Weijie Zhao , Shenglan Li , Denghui Zhang This is my paper
classification cs.IR
keywords financialstatementauditinglargelanguagemodelsbenchmarkfive-stageevaluationaccountingstandardscitationerrordetectionGPT-4synthetictransactiondata
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces a benchmark that pairs real financial tables from S&P 500 companies with synthetic transaction records and injected errors, and it uses a five-stage evaluation to test how well LLMs audit those statements. The central claim is that current LLMs are excellent at the first screening step—they correctly classify both correct and erroneous statements in 100% of cases—but they fail when asked to cite accounting standards, explain errors, and produce corrected tables, with GPT-4 completing a full audit on only 4.1% of single-error statements. If true, this means off-the-shelf LLMs cannot act as independent auditors but could serve as a reliable error-screening front end in an audit workflow, with humans handling the explanation, citation, and revision stages. The paper attributes the failure to missing domain-specific accounting knowledge and weak joint reasoning over tables and text, and it suggests fine-tuning on accounting standards and retrieval-augmented generation as the next steps.

What carries the argument

The central machinery is the five-stage evaluation framework itself, which decomposes auditing into general judgment, error identification, error resolution, standards citation, and financial statement revision, each scored with its own metric (exact match, BertScore, top-K retrieval EM, BLEU, and an overall strict success rate). The pipeline feeds an LLM agent a pair of inputs—a structured textual financial table (from real S&P 500 statements) and a set of historical transaction records (synthetic, GPT-4-generated from few-shot examples written by accounting students)—and supplies a curated database of FASB accounting standards as retrievable memory for the citation stage. Errors are injected in four defined types: missing row, numerical error, redundant row, and misclassification, which lets the benchmark produce 1,484 single-error tables and 371 multi-error tables from 371 real statements. The framework's design makes the boundary visible: each stage inherits errors from the previous one, so a small gap at identification compounds into a near-zero overall success rate.

What would settle it

Run the same five-stage prompts on a test set built from real transaction ledgers and real financial statements whose errors and reference explanations are vetted by professional auditors; if the overall success rate on that set is materially higher than 4.1%, or if professional auditors reject a large share of the GPT-4-generated ground-truth labels, the observed performance boundary is an artifact of the synthetic test bed rather than a fixed LLM limit.

Watch

Extended reading notes

Core claim

The discovery is a performance boundary: in a five-stage audit pipeline, state-of-the-art LLMs achieve perfect or near-perfect performance on binary general judgment (100% EM for both correct and erroneous tables) but the ability to localize and correct errors degrades sharply from there. On tables with a single injected error, GPT-4 attains 0.899 EM on error-type identification and 0.737 EM on pinpointing the problematic row, 0.878 BertScore on error resolution, only 0.262 top-1 EM on citing the correct FASB accounting standard, and 0.783 BLEU on revising the table, for an overall success rate of 4.1%. With multiple errors the numbers fall further. The paper concludes that current LLMs are not reliable enough to execute complete audits and that the bottleneck is a lack of domain-specific accounting knowledge plus weak joint reasoning over tabular and textual data.

Load-bearing premise

The load-bearing premise is that GPT-4-generated synthetic transaction data and human-verified error labels faithfully represent the relationship between a company's transactions and its financial statement line items; if the test bed is too simple or unrealistic, both the perfect detection scores and the low success rates are artifacts of the benchmark rather than true measures of LLM audit skill.

Editorial extensions

If this is right

  • LLMs can be deployed as a first-pass screening tool in audit workflows to flag statements as correct or incorrect with near-perfect reliability, reducing the search space for human auditors.
  • A full audit cannot be delegated to current off-the-shelf LLMs; any automated pipeline must keep a human in the loop for error explanation, standards citation, and table correction.
  • Improvements should target domain-specific accounting knowledge—for example, fine-tuning on FASB standards and audit procedures, or retrieval-augmented generation over a standards knowledge base—rather than general reasoning ability alone.
  • Systems that reason jointly over tables and text, such as multi-modal or table-aware architectures, are a necessary direction because models currently struggle to locate and correct errors even when given the error type.
  • The overall success rate will drop further as statements contain multiple errors, so real-world, error-dense statements will be even harder for LLMs to audit than the benchmark's single-error cases.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The perfect detection score may be partly an artifact of the benchmark's error-injection scheme, because the errors are generated by the same GPT-4 family that is being tested, making them statistically easier to spot than genuine human or real-world misstatements.
  • The paper's boundary suggests a concrete product shape: a human-in-the-loop audit assistant that auto-flags candidate errors and generates draft explanations, with a professional auditor validating citations and edits; this would preserve the paper's demonstrated screening value while working around the citation and revision failures.
  • A testable extension would be to re-run the five-stage evaluation with a standards database that includes the full original FASB text rather than practitioner summaries; if citation accuracy jumps, the bottleneck is retrieval and grounding rather than missing accounting knowledge.
  • The overall success rate metric is strict (all stages must succeed), so even modest per-stage errors quickly compound; in practice, a user might accept a lower bar for earlier stages, meaning the 4.1% figure overstates the gap between LLMs and a usable audit tool.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. This paper introduces a benchmark and five-stage evaluation framework for assessing LLMs as financial statement auditors. The data combine real S&P 500 financial statements (converted to structured text) with GPT-4-synthesized transaction data, and the benchmark includes correct tables, tables with one injected error, and tables with multiple injected errors. The authors evaluate GPT-3.5-Turbo and GPT-4 on general judgment, error identification, error resolution, standards citation, and table revision, plus an overall success rate. They report perfect General Judgment scores (1.000 EM) on correct and erroneous tables, moderate scores for error type identification and entry localization, and poor scores for standards citation (top-1 0.262 for GPT-4) and overall success (4.1% on single-error tables). The paper concludes that current LLMs are suitable as an initial error-screening front end but cannot yet complete full audits.

Significance. If the quantitative results survive closer scrutiny, the paper provides a useful diagnostic tool and a realistic capability ceiling for off-the-shelf LLMs in a domain-specific, table-plus-text reasoning task. The five-stage decomposition is a sensible structure for isolating where automated auditing breaks down, and the use of real financial statement tables is a strength. However, the central quantitative claims currently rest on two fragile pillars: the synthetic transaction data may leak the correct table structure, and the ground-truth explanations, citations, and corrected statements are generated by the same model family being evaluated. Both are fixable, and the later-stage failure results (low citation accuracy, low overall success) are directionally consistent with other domain-expertise findings, so the paper's high-level message is plausible. With added baselines, confidence intervals, and a clear statement that row-aligned labels are removed, the benchmark could be a valuable community resource.

major comments (5)
  1. [Data Creation / Transactions; Appendix: Templates to Generate Synthetic Transaction Data] The transaction-generation prompt explicitly emits row-aligned blocks of the form "[contributing to row 1]: Cash and cash equivalents | $41,862" followed by the transactions that sum to that line item. The manuscript never states that these row labels are stripped from the stored transaction inputs before they are passed to the auditor prompt; the only stripped example appears in the few-shot auditor template, which is not evidence about the actual evaluation inputs. If the stored inputs retain these headings, then a Missing Row, Redundant Row, or Misclassification is detectable by format matching rather than audit reasoning, and the reported 1.000 EM on General Judgment and the high Error Entry scores would be inflated. The authors must either confirm that the stored inputs are stripped or rerun the evaluation with stripped inputs and report the difference; this is load-bearing for the advertised "safe as initial screening" conclusion.
  2. [Errors and Explanations] The ground truth for error resolution, standards citation, and corrected statements is generated by GPT-4 together with the injected errors, and the text states that "since the GPT models are the generator of these errors, it is easy to ensure that the ground truth labels are aligned and accurate." This is a same-family reference standard: GPT-4 is then evaluated against references it produced. A correct but differently phrased resolution will receive a lower BertScore, and a valid alternative FASB citation not present in the single-reference ground truth will count as a miss under Top-1 EM. The paper should provide multiple human-validated reference standards or at minimum an independent audit of the ground-truth labels by accounting professionals, and it should discuss the expected direction of the bias on each metric.
  3. [Evaluation Metrics (Standards Citation)] The standards-citation stage is described only as "an external retriever extracts relevant accounting standards from our standards database," but the retriever is never specified. The Top-1 and Top-5 EM scores depend entirely on the retriever's indexing and ranking; without information about whether it is dense, sparse, or LLM-based, and how the free-text ground-truth citations are mapped to standard indices, the reported citation results (e.g., GPT-4 Top-1 0.262 on single-error tables) cannot be interpreted or reproduced. This component is load-bearing for the later-stage failure claim and must be specified.
  4. [Tables 2 and 3; Limitation 3] The single-error and multiple-error results are averages over 150 random samples, yet Limitation 3 states the evaluation was run on a sample of 300 financial statements; the manuscript does not reconcile these numbers. No confidence intervals or significance tests are reported, so the perfect 1.000 General Judgment scores could be consistent with a non-trivial error rate in the population, and the conclusion that LLMs are safe as a screening front end is not statistically supported. In addition, there are no baseline comparisons (e.g., a rule-based line-item match or a majority-class predictor) against which to judge the detection and identification scores. The authors should clarify the sample size and add variance estimates or per-error-type breakdowns plus at least one simple baseline.
  5. [Evaluation Metrics (Overall Auditing Process)] The overall Success Rate is defined with arbitrary thresholds: all EM scores equal to 1, BertScore above 0.85, and BLEU above 0.99. The BLEU threshold is so strict that a semantically correct table revision with minor formatting or ordering differences will fail, so the reported 4.1% (single-error, GPT-4) success rate conflates auditing ability with exact lexical reproduction of the GPT-4-generated correction. The authors should report results across a range of thresholds or justify the chosen values; otherwise the headline "LLMs cannot complete a full audit" is overstated.
minor comments (5)
  1. [LLMs Effectively Avoid Misclassification of Error-Free Financial Statements] This paragraph refers to "Table 2" for the correct-table results, but the correct-table results are in Table 1; the following paragraph also mislabels the single-error table as Table 3 and the multiple-error table as Table 1. The cross-references need to be corrected.
  2. [Limitation 3 and Table captions] The manuscript is internally inconsistent about the evaluation sample: the table captions say 150 random samples while Limitation 3 says 300 financial statements. Please clarify the exact number of evaluation instances and how the 150 samples relate to the 371 statements and the 1,484 single-error tables.
  3. [Benchmark release] The paper claims to release a benchmark, but no dataset URL, repository, or artifact link is provided in the manuscript. For a benchmark paper, this is a significant presentation omission and should be fixed.
  4. [Standards citation mapping] The evaluation description says the retriever returns a standard's index and compares it with the ground truth, but the ground-truth citations shown in the appendix are free-text FASB quotes such as "FASB ASC 210-10-45-1." The paper should specify how free-text citations are normalized to indices for the EM comparison.
  5. [Typos and wording] There are several typographical issues, including "124 images form from the S&P500 Companies" (Data Creation), "Trasactons" in the appendix prompt, and "uff1a" in the injected example. Additionally, the introduction says LLMs are "sufficient" for identifying errors while the discussion later says they are "insufficient to serve as reliable auditors"; the wording should be made consistent.

Circularity Check

1 steps flagged · score 5.0 of 10

GPT-4 writes the answer key and then takes the exam: the stage 3-5 scores for error resolution, standards citation, and table revision measure agreement with GPT-4-authored ground truth, so the headline 'significant limitations in explaining and citing' claim rests on a same-family reference loop; the detection-stage claims remain structurally independent.

  1. self definitional [Data Creation / Errors and Explanations; Experiments / Evaluation Metrics; Table 2]
    "Along with the erroneous tables, we generate five-stage evaluation ground truth labels simultaneously. Since the GPT models are the generator of these errors, it is easy to ensure that the ground truth labels are aligned and accurate. ... Error resolution is evaluated using BertScore, which assesses the semantic similarity between the generated results and the ground truth. ... The modifications to table entries are compared to the ground truth using BLEU."

    The five-stage ground-truth answers are authored by GPT-4 in the error-injection step: the same template that inserts an error also makes GPT-4 emit the modified statement plus 'Error Resolution', 'Standards Citation', and a corrected statement, and the data-creation section states these labels are generated by the GPT models themselves. The evaluation then scores GPT-4 and GPT-3.5-Turbo by BertScore, Top-K retrieval EM, and BLEU against exactly those GPT-4-authored texts. 'Correct explanation / standard / revision' is therefore defined as GPT-4's own wording; the reported stage 3-5 numbers (BERTScore 0.878, Top-1 citation 0.262, BLEU 0.783) measure self-similarity with the answer-key author, not validated accounting correctness.

full rationale

This paper contains no equations or fitted parameters; the 'derivation chain' is benchmark construction -> evaluation -> conclusions, so circularity can only reside in how the reference standard is defined. One genuine loop is present: the five-stage ground truth is generated by the GPT models ('we generate five-stage evaluation ground truth labels simultaneously'), and the same models are then evaluated against it. For stages 1-2 the labels are objective (a row was deleted, moved, or given a wrong number, and the modified table genuinely conflicts with the transaction data), so the 1.000 EM detection results and the proposed front-end screening role are independent of the loop. For stages 3-5 the reference text is GPT-4's own output, and BertScore / Top-K EM / BLEU measure agreement with that text; the conclusion that LLMs lack domain-specific accounting knowledge is therefore evidenced by a same-family agreement score, not by comparison to FASB sources or independent expert correction. This warrants a moderate score (5): the central boundary claim does not fully reduce, but three of the five stages rest on a self-referential answer key. The skeptical row-header leakage concern is not demonstrable from the manuscript: the transaction-generation template emits '[contributing to row n]: Account | amount' headings, while the auditor-prompt few-shot example shows a flattened transaction list; the stored input format is unstated, so whether detection reduces to format matching cannot be claimed from the text and is flagged here as a data-hygiene risk rather than a circular step. The Limitation section explicitly concedes that 'the transaction data provided in our benchmark is simplified and lacks scalability' and that only 150 random samples per condition were evaluated; these are validity concessions and are weighed as supporting the moderate score rather than raising it. Self-citations (Wang et al. 2023; Yang et al. 2024; Liu et al. 2024a,b; Deng et al. 2024) appear only in the Limitation section as suggestions for future work and are not load-bearing, so they do not contribute to the score. The strict overall-success threshold (all EM=1, BERTScore>0.85, BLEU>0.99) is an explicit design choice, not a hidden fit.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

No numerical parameters are fitted in the traditional sense, but the evaluation design contains hand-chosen thresholds and a sample-size choice that act as free parameters. The central results additionally rest on the realism of GPT-4-generated synthetic transactions and on the validity of GPT-4-generated ground truth for the later audit stages.

free parameters (2)
  • Overall success thresholds = EM=1, BertScore>0.85, BLEU>0.99
    Hand-chosen thresholds for declaring a complete audit successful; changing them changes the headline 2-4% success rates.
  • Evaluation sample size = 150 random samples per condition
    Results are averages over a 150-sample subset; the authors flag in the Limitation section that this may not capture full-dataset performance.
assumptions (4)
  • domain assumption GPT-4-generated synthetic transaction data, prompted by accounting students, faithfully aligns with the real company financial statement line items and represents realistic audit evidence.
    Invoked in Data Creation / Transactions; acknowledged as simplified in Limitation section 1.
  • domain assumption The four error types (missing row, numerical error, redundant row, misclassification) are a sufficient taxonomy for financial statement auditing errors.
    Introduced in Data Creation / Errors and Explanations; no evidence that other error types, such as conceptual classification errors across statements, are covered.
  • ad hoc to paper GPT-4-generated ground truth for error resolution, standards citation, and corrected statements is treated as correct without validation against professional auditors.
    Data Creation states ground truth labels are generated simultaneously by GPT-4; no independent auditor verification is described.
  • domain assumption The manually curated summaries of FASB standards used as the retrieval knowledge base are accurate and complete.
    Data Creation states professional accounting practitioners manually summarized standards, but the summaries are not shown or validated in the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Automating Financial Statement Audits with Large Language Models." pith.science (2026). https://pith.science/paper/ILI7MAI5

@misc{pith2026250617282,
  author       = {Pith},
  title        = {Pith review of: Automating Financial Statement Audits with Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ILI7MAI5}},
  note         = {Machine review of arXiv:2506.17282}
}
read the original abstract

Financial statement auditing is essential for stakeholders to understand a company's financial health, yet current manual processes are inefficient and error-prone. Even with extensive verification procedures, auditors frequently miss errors, leading to inaccurate financial statements that fail to meet stakeholder expectations for transparency and reliability. To this end, we harness large language models (LLMs) to automate financial statement auditing and rigorously assess their capabilities, providing insights on their performance boundaries in the scenario of automated auditing. Our work introduces a comprehensive benchmark using a curated dataset combining real-world financial tables with synthesized transaction data. In the benchmark, we developed a rigorous five-stage evaluation framework to assess LLMs' auditing capabilities. The benchmark also challenges models to map specific financial statement errors to corresponding violations of accounting standards, simulating real-world auditing scenarios through test cases. Our testing reveals that current state-of-the-art LLMs successfully identify financial statement errors when given historical transaction data. However, these models demonstrate significant limitations in explaining detected errors and citing relevant accounting standards. Furthermore, LLMs struggle to execute complete audits and make necessary financial statement revisions. These findings highlight a critical gap in LLMs' domain-specific accounting knowledge. Future research must focus on enhancing LLMs' understanding of auditing principles and procedures. Our benchmark and evaluation framework establish a foundation for developing more effective automated auditing tools that will substantially improve the accuracy and efficiency of real-world financial statement auditing.

Figures

Figures reproduced from arXiv: 2506.17282 by the authors.

Figure 1
Figure 1. Our Automated Financial Statement Auditing pipeline: An LLM-based agent is provided with (a) The history transaction [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. APEX-Accounting

    cs.CL 2026-07 conditional novelty 7.0 of 10

    Frontier agents top out at 56.4% Mean Criteria@3 and ≤2.6% Pass^8 on expert-authored multi-file accounting close tasks across ten synthetic company worlds.

  2. ISACL: Internal State Analyzer for Copyrighted Training Data Leakage

    cs.CL 2025-08 conditional novelty 5.0 of 10

    An MLP trained on LLM internal states predicts Rouge-L-defined literal copying leakage with high accuracy, but not paraphrase-level leakage.

Reference graph

Works this paper leans on

24 extracted references · 18 canonical work pages · cited by 2 Pith papers

  1. [1]

    Berger, A.; Hillebrand, L.; Leonhard, D.; Deu er, T.; De Oliveira, T. B. F.; Dilmaghani, T.; Khaled, M.; Kliem, B.; Loitz, R.; Bauckhage, C.; et al. 2023. Towards automated regulatory compliance verification in financial auditing with large language models. In 2023 IEEE International Conference on Big Data (BigData), 4626--4635. IEEE

    work page 2023

  2. [2]

    E.; Al-Awadhi, A.; Gullvist, B.; Brown-Liburd, H.; Teeter, R.; Warren Jr, J

    Byrnes, P. E.; Al-Awadhi, A.; Gullvist, B.; Brown-Liburd, H.; Teeter, R.; Warren Jr, J. D.; and Vasarhelyi, M. 2018. Evolution of auditing: From the traditional approach to the future audit. In Continuous auditing: Theory and application, 285--297. Emerald Publishing Limited

    work page 2018

  3. [3]

    Cao, M.; Chychyla, R.; and Stewart, T. 2015. Big data analytics in financial statement audits. Accounting horizons, 29(2): 423--429

    work page 2015

  4. [4]

    Cho, J. 2022. How Do Auditors Respond to Labor Investment Inefficiency? The Journal of the Korea Contents Association, 22(2): 593--604

    work page 2022

  5. [5]

    Debreceny, R.; Lee, S.-L.; Neo, W.; and Toh, J. S. 2005. Employing generalized audit software in the financial services sector: Challenges and opportunities. Managerial Auditing Journal, 20(6): 605--618

    work page 2005

  6. [6]

    Deng, J.; Wei, Z.; Pang, L.; Ding, H.; Shen, H.; and Cheng, X. 2024. UnKE: Unstructured Knowledge Editing in Large Language Models. arXiv preprint arXiv:2405.15349

    arXiv 2024

  7. [7]

    J.; Akresh, A

    Elder, R. J.; Akresh, A. D.; Glover, S. M.; Higgs, J. L.; and Liljegren, J. 2013. Audit sampling research: A synthesis and implications for future research. Auditing: A Journal of Practice & Theory, 32(Supplement 1): 99--129

    work page 2013

  8. [8]

    Hillebrand, L.; Berger, A.; Deu er, T.; Dilmaghani, T.; Khaled, M.; Kliem, B.; Loitz, R.; Pielka, M.; Leonhard, D.; Bauckhage, C.; et al. 2023. Improving zero-shot text matching for financial auditing with large language models. In Proceedings of the ACM Symposium on Document Engineering 2023, 1--4

    work page 2023

Show all 24 references
  1. [9]

    Hu, L.; Liu, Z.; Zhao, Z.; Hou, L.; Nie, L.; and Li, J. 2023. A survey of knowledge enhanced pre-trained language models. IEEE Transactions on Knowledge and Data Engineering

    2023

  2. [10]

    Li, J.; and Sun, Z. 2023. Government audit, employee efficiency and labor cost stickiness. Plos one, 18(9): e0291014

    2023

  3. [11]

    Liu, J.; Ai, L.; Liu, Z.; Karisani, P.; Hui, Z.; Fung, M.; Nakov, P.; Hirschberg, J.; and Ji, H. 2024 a . PropaInsight: Toward Deeper Understanding of Propaganda in Terms of Techniques, Appeals, and Intent. arXiv preprint arXiv:2409.18997

    2024 arXiv

  4. [12]

    Liu, J.; Yu, P.; Zhang, Y.; Li, S.; Zhang, Z.; Sarikaya, R.; Small, K.; and Ji, H. 2024 b . EVEDIT : Event-based Knowledge Editing for Deterministic Knowledge Propagation. In Al-Onaizan, Y.; Bansal, M.; and Chen, Y.-N., eds., Proceedings of the 2024 Conference on Empirical Met...

    2024

  5. [13]

    Nigrini, M. J. 2017. Audit sampling using Benford's Law: A review of the literature with some new perspectives. Journal of emerging technologies in accounting, 14(2): 29--46

    2017

  6. [14]

    Popovi \'c , S.; To s kovi \'c , J.; Majstorovi \'c , A.; Brkanli \'c , S.; and Kati \'c , A. 2015. THE IMPORTANCE OF CONTINUOUS AUDIT OF FINANCIAL STATEMENTS OF THE COMPANY OF COUNTRIES JOINING THE EU. Annals of'Constantin Brancusi'University of Targu-Jiu. Economy Series/Anal...

    2015

  7. [15]

    Sirikulvadhana, S. 2002. Data mining as a financial auditing tool. Ph.D. thesis, Citeseer

    2002

  8. [16]

    Stewart, T. R. 2015. Data analytics for financial statement audits. Audit Analytics, 105

    2015

  9. [17]

    Sun, Y.; Li, J.; Lu, M.; and Guo, Z. 2024. Study of the Impact of the Big Data Era on Accounting and Auditing. arXiv preprint arXiv:2403.07180

    2024 arXiv

  10. [18]

    A.; Yahya, S.; and Siam, W

    Toumeh, A. A.; Yahya, S.; and Siam, W. Z. 2018. Expectations gap between auditors and user of financial statements in the audit process: an auditors' perspective. Asia-Pacific Management Accounting Journal (APMAJ), 13(3): 79--107

    2018

  11. [19]

    Vinatoru, S.; and Domnisoru, S. 2016. USING DATA MINING IN FINANCIAL AUDIT. In 3rd International Multidisciplinary Scientific Conference on Social Sciences and Arts SGEM 2016, 677--684

    2016

  12. [20]

    Wang, X.; Wang, Z.; Liu, J.; Chen, Y.; Yuan, L.; Peng, H.; and Ji, H. 2023. Mint: Evaluating llms in multi-turn interaction with tools and language feedback. arXiv preprint arXiv:2309.10691

    2023 arXiv

  13. [21]

    V.; Zhou, D.; et al

    Wei, J.; Wang, X.; Schuurmans, D.; Bosma, M.; Xia, F.; Chi, E.; Le, Q. V.; Zhou, D.; et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35: 24824--24837

    2022

  14. [22]

    R.; Li, S.; Huang, Z.; Cao, X.; Wang, X.; Wang, Y.; et al

    Yang, K.; Liu, J.; Wu, J.; Yang, C.; Fung, Y. R.; Li, S.; Huang, Z.; Cao, X.; Wang, X.; Wang, Y.; et al. 2024. If llm is the wizard, then code is the wand: A survey on how code empowers large language models to serve as intelligent agents. arXiv preprint arXiv:2401.00812

    2024 arXiv

  15. [23]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  16. [24]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.