Pith. sign in

REVIEW 5 major objections 6 minor 24 references

TeXFix-Bench: An Empirically Grounded Multi-Format Benchmark for LLM-Based Document Source Repair

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper argues that evaluating LLM document repair on arbitrary injected typos overstates performance, and that repair benchmarks should be grounded in a mined taxonomy of real human crashes and judged under both compile and…

desk verdict A genuinely useful LaTeX repair benchmark with a strong two-oracle protocol; the multi-format extension is thinner than the title suggests but openly scoped. read the letter →

arxiv 2608.07617 v1 pith:RXQWU3WP submitted 2026-08-07 cs.AI

Prajwal S. Venkateshmurthy This is my paper · ORCID
classification cs.AI
keywords automatedprogramrepairLaTeXTypstMarkdownmutationtestingfaulttaxonomyLLMevaluationtwo-oraclebenchmark
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

This paper argues that evaluating LLM document repair on arbitrary injected typos overstates performance, and that repair benchmarks should be grounded in a mined taxonomy of real human crashes and judged under both compile and content-restoration oracles. It mines 168 real LaTeX compilation crashes from a Q&A forum, commit histories, and package documentation, codes them into 18 fault categories, and turns those categories into a 48-operator mutation library called DocMut. On matched seeds, repairs of DocMut faults are 5.6–9.2 percentage points harder than repairs of pattern-based mutations across three model families, and 88 real human crashes are harder still, with a 67.0% repair success rate. The resulting 10,437-instance benchmark is scored twice: engine compile success, and a text-restoration similarity score. The restoration oracle shows that among repairs that compile, 13.6–18.5% materially alter document text, so compile-only leaderboards overstate repair quality.

What carries the argument

The load-bearing machinery is the mined fault taxonomy plus its instantiation and scoring. The taxonomy is an 18-category scheme derived by dual open coding of 168 verified localized hard crashes in LaTeX, with a documented reliability analysis. DocMut is the 48-operator, format-aware mutation library (25 LaTeX, 15 Typst, 8 Markdown) that instantiates the taxonomy into concrete broken sources. The two-oracle protocol carries the evaluation: a pinned engine gate (Tectonic for LaTeX, Typst engine for Typst, pandoc for Markdown) decides compile success, and a restoration oracle computes token-LCS Dice similarity between the extracted text of the candidate and the golden document to measure whether content survived.

What would settle it

Collect a separate set of real human Typst compilation crashes from public forums and issue trackers, classify them against the 18 LaTeX-derived categories, and repair them with the same models under the same protocol; if the real Typst fault distribution is dominated by categories outside the analogical mapping, or if models repair real Typst crashes at rates far above or below DocMut-generated Typst instances, the benchmark's cross-format validity claim would be falsified.

Watch

Extended reading notes

Core claim

The central claim is that how faults are injected into a document-repair benchmark changes what the benchmark measures: pattern-based edits inflated repair scores by up to ten points, while faults modeled on real human crashes were harder for every model family tested, and genuinely mined human crashes were harder still. The paper further claims that compile success is an insufficient quality oracle, since a large fraction of compiling repairs delete or rewrite substantial document content, and that the model ranking by content restoration diverges from the ranking by compile rate. It presents this as an end-to-end loop: mine human faults, derive a taxonomy, instantiate it as a mutation library, validate that the library produces realistic difficulty, then run a large multi-model evaluation under two oracles.

Load-bearing premise

The cross-format claims assume that the 18 fault categories mined from LaTeX crashes transfer to Typst and Markdown, but the Typst and Markdown DocMut operators were designed by analogy rather than mined; if real faults in those formats follow a different distribution, the benchmark's cross-format difficulty rankings are not representative.

Editorial extensions

If this is right

  • Document-repair leaderboards should report a content-fidelity statistic alongside compile success, because compile-only rankings can reward aggressive full-rewrite behavior that destroys content.
  • Benchmark builders for new markup or programming domains should mine human faults before designing mutation operators, since operator selection changes measured difficulty by roughly five to ten points under fixed models and seeds.
  • The Typst gap, which appears across nearly all models, indicates that newer syntax and less training exposure make repairs markedly harder, so cross-language difficulty should be reported per format rather than averaged away.
  • Provider-pinned routing and a per-request provider ledger make delivery failures attributable to serving stacks, meaning deployment-facing rates should be separated from conditional model-skill rates.
  • The zero-shot protocol without diagnostics gives lower bounds on assisted repair performance; supplying compiler errors as input is a planned extension that would likely raise the reported success rates.

Reading between the lines

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

  • If the restoration-oracle result generalizes, compile-only leaderboards may systematically favor models that patch by deletion, so product builders should weight content fidelity when choosing a repair model.
  • Mining Typst and Markdown human crashes directly would test the analogical extension of the LaTeX taxonomy; the 18 LaTeX categories may not cover Typst's scripting and set-rule faults or Markdown's rendering semantics.
  • The 67.0% success rate on real human crashes suggests realistic repair accuracy is well below the compile rates seen on synthetic sets, so future benchmark versions could include permanently held-out real-crash instances as a calibration check.
  • Extending the restoration oracle from extracted-text similarity to layout or visual similarity could sharpen the measurement of content fidelity, particularly for Typst and Markdown where text extraction is only a proxy.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper introduces TeXFix-Bench, a benchmark for full-source document repair by LLMs across LaTeX, Typst, and Markdown. It first mines 168 real localized LaTeX hard crashes from TeX.SE, GitHub commits, and package documentation, and derives an 18-category fault taxonomy through dual open coding. The taxonomy is instantiated as DocMut, a 48-operator AST-aware mutation library (25 LaTeX, 15 Typst, 8 Markdown). Using a three-model cross-benchmark on matched seeds, the paper reports that DocMut faults are 5.6–9.2 percentage points harder to repair than pattern-based mutations, and a real-error case study on 88 human crashes reports 67.0% repair success, bracketing the synthetic sets from below. The paper then constructs a 10,437-instance benchmark from 743 openly licensed seeds, evaluates seven LLMs under a fixed zero-shot protocol with provider-pinned routing, and reports a 27.5-point intention-to-treat compile spread. A restoration oracle over 28,129 compiling repairs shows that 13.6–18.5% of compiling repairs materially alter document text, and that restoration rank diverges from compile rank. The benchmark, taxonomy, operator library, and campaign artifacts are released.

Significance. If the central claims hold, the paper supplies a valuable and much-needed evaluation standard for document-repair LLMs. Its strengths include unusually transparent reporting: the low inter-rater kappa is disclosed and analyzed, intention-to-treat accounting is retained, provider-pinned routing is documented with a per-request ledger, the balanced 6,613×7 model matrix is used for primary claims, and the 14.3% unscorable restoration rows are explicitly accounted for rather than hidden. The restoration oracle is a genuinely useful methodological contribution to the compile-only leaderboard debate. The release of the taxonomy, DocMut, and campaign artifacts makes the work reproducible. However, the multi-format identity of the benchmark rests on an untested analogy transfer from LaTeX to Typst and Markdown, and the real-error case study is partly circular because it reuses the corpus from which the taxonomy was derived. These issues temper the strength of the headline cross-format and validation claims.

major comments (5)
  1. [§4, §8.2, §11] The Typst and Markdown operators are not derived from mined human faults: Section 4 states they 'share the adequacy framework and are designed by analogy' from the LaTeX-mined taxonomy, and Section 11 lists this as an external-validity threat. Yet the abstract and contributions present the benchmark as 'multi-format' and 'empirically grounded,' and Section 8.2's headline that 'Typst is markedly harder' is computed on instances generated by these analogue operators. If real Typst or Markdown hard failures concentrate in categories outside the 18 LaTeX categories, the format-difficulty ranking and the restoration-oracle results for those formats are properties of the artificial operator distribution, not of real repair workloads. This is load-bearing for the benchmark's identity as a multi-format standard. A concrete fix would be to mine a modest sample of real Typst and Markdown hard crashes, compare their category coverage against the 18 LaTeX categories, and either validate the transfer or restrict the 'empirically grounded' claim to LaTeX.
  2. [§5.3] The real-error case study is used as evidence that 'human faults are harder still' and that taxonomy-guided injection moves synthetic evaluation toward real difficulty. However, the 88 evaluated cases are a stratified sample of the same 168 faults from which the taxonomy and, through the taxonomy-refinement loop in Section 4, the DocMut operators were derived. The ordering pattern > DocMut > real is therefore not an independent validation of the taxonomy's adequacy: the operators were refined precisely on these faults. An independent test would apply DocMut to a held-out set of newly mined hard crashes (or at least to faults excluded from the taxonomy derivation) and compare repair success on those. The current framing overstates the evidential value of the case study.
  3. [§3.2] The inter-rater reliability of the taxonomy is kappa=0.34 under free-form open coding, which is 'fair' and below conventional thresholds for instrument reliability. The paper explains the granularity asymmetry and adds an operationalizability check (kappa=0.64) against Annotator 1's labels, but that check is a deterministic classifier encoding the resolution rules, not a second human annotation with the codebook. Because the entire benchmark rests on the claim that the taxonomy is 'empirically grounded,' the absence of a closed-codebook human reliability round remains a load-bearing gap. A second-round human annotation on a subset of the 168 faults would materially strengthen the taxonomy's credibility.
  4. [§9.1] The claim that 'the model with the lowest compile rate restores content best among its successes' (Qwen) is based on mean S_pdf (0.971 vs. Grok's 0.966), a difference of 0.005. The <0.95 tail share, however, is 13.9% for Qwen and 13.6% for Grok, so on the tail metric Qwen is not best. The paper should report both moments with uncertainty (e.g., bootstrapped CIs) or with a pre-specified restoration ranking metric, and should soften the claim if the divergence is not consistent across summary statistics. The broader point that 13.6–18.5% of compiling repairs materially alter text is unaffected.
  5. [§7.3, §8.2] The format-difficulty results in Section 8.2 are computed on the 'full recorded set,' which has uneven per-model coverage (6,613–7,007 instances per model), while the stated design is that all primary claims are verified on the balanced 6,613×7 matrix. The text does not show the format ranking on the balanced matrix, so the claim that the balanced matrix 'confirms all rankings' is not demonstrated for RQ2. Since the full recorded set mixes different model samples across instances, the aggregate format rates could be sensitive to which models attempted which instances. Please report the format ranking on the balanced matrix.
minor comments (6)
  1. [§5.1] Please state explicitly that the cross-benchmark is LaTeX-only (the fault patterns listed are LaTeX-specific), because the current text and Table 2 do not identify the format and the conclusion is subsequently generalized to all DocMut operators.
  2. [§6.2, Appendix A] The text says the engine-gate ledger covers 38 operators and 39 have nonzero instance support, but Table 9 lists 36 codes and omits the mentioned TEX-MTH-OPS single-instance operator; please reconcile these counts.
  3. [Abstract, §1] The parenthetical qualification about Typst/Markdown being extended by analogy is good, but it appears only in the abstract; consider stating it prominently in Section 1 and in the contributions list, where the multi-format claim is made.
  4. [§8.2] The statement that the Typst gap appears across nearly all models would be easier to verify if the per-model Typst CRitt rates were tabulated alongside the span.
  5. [§9] Reporting the fraction of candidates with S_pdf exactly 1.0 would help quantify how the median of 1.0 coexists with the 13.6–18.5% tail below 0.95.
  6. [References] Reference [10] appears malformed ('ACM Comput. Surv.. Extended bibliography'); please clean up the reference entry.

Circularity Check

1 steps flagged · score 4.0 of 10

Real-error 'validation' reuses the 168-fault corpus that generated the taxonomy and refined DocMut operators, making the human-faults-are-harder claim in-sample; the rest of the benchmark chain is independent.

  1. fitted input called prediction [Section 5.3 (Real-error case study); Section 4 (Taxonomy-driven refinement)]
    "For ecological validity we also repaired mined human hard crashes (DeepSeek V4 Pro). From the 168 localized faults we stratified 100 cases. Excluding custom-class documents and one non-failing sandbox case leaves 88 evaluated cases (dummy assets provisioned, .tex unmodified). Compile success is 67.0% (59/88)... the ordering (pattern > DocMut > real) indicates taxonomy-guided injection moves synthetic evaluation toward real difficulty, not past it."

    These 88 'real human crashes' are a stratified sample of the same 168 localized faults from which the taxonomy was mined (Section 3). Section 4 shows the same corpus drove operator design: 'The study validates and refines them: (1) BraceDrop split into runaway vs. extra-brace subpatterns, (2) PackageDrop refined..., (3) soft FontSwap deprecated for hard-crash evaluation.' The case study is therefore an in-sample validation: operators were refined against these very faults, so the 'pattern > DocMut > real' ordering partly reflects fitting to the evaluation corpus, not an independent ecological check. The 'Case study only' caveat narrows but does not remove the overlap.

full rationale

TeXFix-Bench's main measurement chain—mining 168 LaTeX crashes, coding them into 18 categories, instantiating DocMut, comparing DocMut vs. pattern-based injection on matched seeds, and running the 7-model/10,437-instance campaign with compile and restoration oracles—is largely self-contained and not circular. The model rankings, Typst difficulty gap, and restoration-oracle findings are genuine empirical outputs of the released benchmark. The one substantive circularity is the real-error case study (Section 5.3): its 88 cases are drawn from the same 168-fault corpus used to derive the taxonomy and refine operators (Section 4), so the claim that human faults are harder and that DocMut moves 'toward real difficulty' is an in-sample validation, not independent ecological evidence. The Typst/Markdown operators are explicitly 'designed by analogy' from the LaTeX taxonomy (Sections 4, 11); this is a disclosed external-validity threat rather than a circular derivation, and the paper does not hide it. Self-citations to DocMut and the v0.3 pilot are not load-bearing for the new results. Overall, the benchmark's central content is independent; the circularity is partial and localized to one validation claim.

Assumptions & free parameters 3 free parameters · 4 assumptions · 3 invented entities

The central measurement apparatus rests on a small number of modeling choices: the scope definition (localized hard crashes), the analogy extension to Typst and Markdown, the restoration similarity proxy, and the balanced-matrix representability. The paper discloses most of these as threats to validity; the ledger itemizes them.

free parameters (3)
  • Restoration similarity threshold tau = 0.95
    Hand-chosen cutoff defining 'materially alter document text'. The paper checks robustness at 0.90 and 0.99, so sensitivity is addressed, but the primary reported shares use 0.95.
  • Completion cap for model output = 4096 tokens
    Protocol choice that systematically truncates long document rewrites; the 16k ablation partially corrects this, but the headline CRitt numbers embed the original cap.
  • Timeline timeout for compile gate = 30 seconds
    Implementation detail of the engine gate that could interact with slow documents, though unlikely to bias central comparisons.
assumptions (4)
  • domain assumption Localized hard crashes under a pinned engine (Tectonic) are the right unit of repair evaluation.
    Section 3 states the scope excludes multi-hunk reconfigurations, warning-only diagnostics, and multi-file asset puzzles; the entire benchmark validity rests on this scope choice.
  • ad hoc to paper Typst and Markdown faults share the LaTeX-mined taxonomy by analogy.
    Section 4: 'the taxonomy is mined for LaTeX, and Typst and Markdown operators extend it by analogy'. No mined fault data supports the Typst and Markdown operators, so cross-format difficulty claims depend on this analogy.
  • domain assumption Token-LCS Dice similarity between extracted text of golden and candidate approximates content preservation.
    Section 9 defines S_pdf; it ignores layout, Unicode normalization, and ligatures, and the Markdown restoration uses HTML similarity, a different construct, as noted in Section 11.
  • ad hoc to paper The balanced 6,613x7 matrix is representative of the full 10,437-instance freeze.
    Section 7.3: incomplete cells are unfinished and primary claims are stated on the balanced matrix. If API delivery failures correlate with instance difficulty, this assumption fails; the paper reports at most 0.6 pp full-set ITT difference.
invented entities (3)
  • TeXFix-Bench v0.4 benchmark independent evidence
    purpose: Standardized evaluation instrument for LLM-based document repair across LaTeX, Typst, and Markdown.
    The 10,437-instance freeze and all artifacts are publicly released, so external groups can run the benchmark themselves.
  • DocMut mutation operator library independent evidence
    purpose: Instantiates the taxonomy as 48 AST-aware fault-injection operators.
    MIT-licensed and versioned at v0.2.0 on GitHub, making it independently usable and checkable.
  • Restoration oracle S_pdf (token-LCS Dice similarity)
    purpose: A second metric for content fidelity beyond compile success.
    The metric is defined inside the paper; it has not been independently calibrated against human judgments of content preservation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TeXFix-Bench: An Empirically Grounded Multi-Format Benchmark for LLM-Based Document Source Repair." pith.science (2026). https://pith.science/paper/RXQWU3WP

@misc{pith2026260807617,
  author       = {Pith},
  title        = {Pith review of: TeXFix-Bench: An Empirically Grounded Multi-Format Benchmark for LLM-Based Document Source Repair},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RXQWU3WP}},
  note         = {Machine review of arXiv:2608.07617}
}
abstract

Scientific and technical writing depends on markup sources that must compile: LaTeX, Typst, and Markdown pipelines fail on missing delimiters, mismatched environments, broken imports, or package conflicts. Existing document-repair evaluations inject faults with ad-hoc edits that lack an empirical fault model. We present TeXFix-Bench, a multi-format benchmark for LLM-based full-source document repair grounded in a mined fault taxonomy. A Grounded-Theory study of localized hard-crash LaTeX faults from TeX Stack Exchange, GitHub commits, and package documentation (168 verified faults, dual open coding at $\kappa$=0.34) yields an 18-category taxonomy instantiated as DocMut: 48 AST-aware operators across three formats. A three-model cross-benchmark shows DocMut faults are 5.6-9.2 pp harder to repair than pattern-based mutations on the same seeds, and a real-error case study (88 mined human crashes, 67.0% repair success) brackets both synthetic sets from below. We construct 10,437 instances from 743 openly licensed seeds and evaluate seven LLMs under a fixed zero-shot protocol with provider-pinned routing, collecting 48,651 attempts at about USD 200 total inference cost. A complete 6,613-instance x 7-model balanced matrix confirms all rankings. A pinned engine gate yields a 27.5-point intention-to-treat compile spread (56.7-84.2%). Typst is markedly harder than LaTeX and Markdown. A restoration oracle over 28,129 compiling repairs shows that 13.6-18.5% of compiling repairs materially alter document text, and restoration rank diverges from compile rank: the model with the lowest compile rate restores content best among its successes. Compile success alone overstates repair quality. We release the taxonomy, DocMut, and all campaign artifacts.

Figures

Figures reproduced from arXiv: 2608.07617 by the authors.

Figure 1
Figure 1. End-to-end pipeline: mined faults ground the taxonomy, the taxonomy grounds the operators, and the resulting [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Frequency of resolved axial categories among the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Deployment-facing CRitt vs. conditional CRapi on the balanced matrix. High conditional skill with low delivery (Qwen, GLM, DS-Pro) yields weak end-to-end rates. Equivalent-mutant rate Fraction of API-ok candidates whose SHA-256 equals the golden (exact reverts). Restoration 𝑆pdf Token-LCS Dice between candidate and golden pdftotext output for compiling non-revert candidates (Section 9). We report rates with explicit… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Compile rate vs. content restoration. The two or [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 16 canonical work pages

  1. [1]

    Virginia Braun and Victoria Clarke. 2006. Using thematic analysis in psychology. Qualitative Research in Psychology3, 2 (2006), 77–101

    2006

  2. [2]

    Alex Groce et al . 2023. Universal Mutator. https://github.com/agroce/ universalmutator

    work page 2023

  3. [3]

    Rahul Gupta, Soham Pal, Aditya Kanade, and Shirish Shevade. 2017. DeepFix: Fixing Common C Language Errors by Deep Learning. InProc. AAAI. 1345–1351

    work page 2017

  4. [4]

    Junyi Hou et al. 2026. PaperDebugger: A Plugin-Based Multi-Agent System for In-Editor Academic Writing. InCompanion Proc. ACM Web Conf.144–147

    work page 2026

  5. [5]

    Yue Jia and Mark Harman. 2011. An Analysis and Survey of the Development of Mutation Testing.IEEE Trans. Softw. Eng.37, 5 (2011), 649–678. https: //doi.org/10.1109/TSE.2010.62

    doi:10.1109/tse.2010.62 2011

  6. [6]

    Jimenez et al

    Carlos E. Jimenez et al. 2024. SWE-bench: Can Language Models Resolve Real- World GitHub Issues?. InProc. ICLR

    work page 2024

  7. [7]

    René Just, Darioush Jalali, and Michael D. Ernst. 2014. Defects4J: A Database of Existing Faults to Enable Controlled Testing Studies for Java Programs. InProc. ISSTA. https://doi.org/10.1145/2610384.2628055

    arXiv 2014

  8. [8]

    Richard Landis and Gary G

    J. Richard Landis and Gary G. Koch. 1977. The measurement of observer agree- ment for categorical data.Biometrics33, 1 (1977), 159–174

    1977

Show all 24 references
  1. [9]

    Claire Le Goues et al . 2015. The ManyBugs and IntroClass Benchmarks for Automated Repair of C Programs.IEEE Trans. Softw. Eng.41, 12 (2015), 1236–

    2015

  2. [10]

    Martin Monperrus. 2018. Automatic Software Repair: A Bibliography. ACM Comput. Surv.. Extended bibliography of APR

    2018

  3. [11]

    Jefferson Offutt, Ammei Lee, Gregg Rothermel, Roland H

    A. Jefferson Offutt, Ammei Lee, Gregg Rothermel, Roland H. Untch, and Christian Zapf. 1996. An experimental determination of sufficient mutant operators.ACM Transactions on Software Engineering and Methodology5, 2 (1996), 99–118

    1996

  4. [12]

    OpenRouter. 2026. OpenRouter Unified LLM API. https://openrouter.ai/

    2026

  5. [13]

    Mike Papadakis, Marinos Kintis, Jie Zhang, Yue Jia, Yves Le Traon, and Mark Harman. 2019. Mutation Testing Advances: An Analysis and Survey.Adv. Comput.112 (2019), 275–378

    2019

  6. [14]

    Smith, Earl T

    Edward K. Smith, Earl T. Barr, Claire Le Goues, and Yuriy Brun. 2015. Is the Cure Worse Than the Disease? Overfitting in Automated Program Repair. In Proc. ESEC/FSE. https://doi.org/10.1145/2786805.2786825

    2015

  7. [15]

    Jirong Tan and Manuel Rigger. 2024. Inconsistencies in TeX-Produced Documents. InProc. ISSTA. 1415–1427. https://doi.org/10.1145/3650212.3680370

    2024

  8. [16]

    Venkateshmurthy

    Prajwal S. Venkateshmurthy. 2026. DocMut: Document Mutation Library. https: //github.com/prajwal-svm/docmut. v0.2.0, MIT

    2026

  9. [17]

    Venkateshmurthy

    Prajwal S. Venkateshmurthy. 2026. TeXFix-Bench v0.3: An Audited Synthetic Benchmark for Single-Fault LaTeX Compilation Repair. Zenodo. https://doi. org/10.5281/zenodo.21764870

    2026 doi

  10. [18]

    Viera and Joanne M

    Anthony J. Viera and Joanne M. Garrett. 2005. Understanding interobserver agreement: The kappa statistic.Family Medicine37, 5 (2005), 360–363

    2005

  11. [19]

    Cheng Wang, Lin Fu, Zhiyuan Kuang, and Yu Zhao. 2026. TexOCR: Advancing Document OCR Models for Compilable Page-to-LaTeX Reconstruction. InProc. ACL. 35821–35845

    2026

  12. [20]

    Haoran Wen et al. 2024. OverleafCopilot: Empowering Academic Writing in Overleaf with Large Language Models. arXiv:2403.09733

    2024 arXiv

  13. [21]

    Chunqiu Steven Xia and Lingming Zhang. 2023. Keep the Conversation Go- ing: Fixing 162 out of 337 bugs for $0.42 each using ChatGPT. InProc. ISSTA. Conversational APR / ChatRepair line of work

    2023

  14. [22]

    Michihiro Yasunaga and Percy Liang. 2020. Graph-Based, Self-Supervised Pro- gram Repair from Diagnostic Feedback. InProc. ICML. 10799–10808. TeXFix-Bench: An Empirically Grounded Multi-Format Benchmark for LLM-Based Document Source Repair Table 9: Operators with support in the...

    2020

  15. [23]

    Fengmin Zhu and Fei He. 2022. EqFix: Fixing LaTeX Equation Errors by Examples. InSETTA (LNCS, Vol. 13649). Springer, 106–124. A Operator Codes Observed in the Freeze Table 9 lists operator codes with nonzero instance counts in the v0.4 freeze. Full rationales ship with theDocM...

    2022

  16. [1256]

    https://doi.org/10.1109/TSE.2015.2454513

    2015

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.