Pith. sign in

REVIEW 4 major objections 6 minor 59 references

SHARE: An SLM-based Hierarchical Action CorREction Assistant for Text-to-SQL

T0 review · 4 major / 6 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 SHARE claims that a pipeline of three small language models, each under 8B parameters, can localize and repair SQL errors more precisely than the large generator model can on its own, improving GPT-4o's execution accuracy by 14.80% on…

desk verdict A solid empirical paper on SLM-assisted self-correction for text-to-SQL; the central claim holds up, but the unmeasured trajectory fidelity and missing error bars keep it from being definitive. read the letter →

arxiv 2506.00391 v1 pith:7WO7NIIE submitted 2025-05-31 cs.CL

Ge Qu , Jinyang Li , Bowen Qin , Xiaolong Li , Nan Huo , Chenhao Ma , Reynold Cheng This is my paper · ORCID
classification cs.CL
keywords text-to-SQLself-correctionsmalllanguagemodelsactiontrajectoryschemalinkinglogicoptimizationhierarchicalself-evolutionexecutionaccuracy
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

SHARE is an attempt to make LLM self-correction for text-to-SQL both more accurate and far cheaper by replacing the recursive self-calls of a large model with a one-pass pipeline of three small language models, each under 8 billion parameters. The paper's central claim is that converting a declarative SQL query into a stepwise "action trajectory" of pandas-like operations exposes the query's reasoning path, so that errors can be localized and repaired more precisely than the LLM can do on its own. In experiments with GPT-4o as the generator, SHARE-8B raises execution accuracy by a relative 14.80% on BIRD and 11.41% on SPIDER in a single correction round, at roughly one-tenth the inference cost of the cheapest compared self-correction method. If this holds, it means small, privacy-friendly models can serve as general correction assistants without repeated expensive LLM calls.

What carries the argument

The central object is the action trajectory: a declarative SQL query decomposed into a sequence of pandas-like function calls (where, select, groupby, orderby, limit, distinct, union, plus aggregation and operator actions) that expose the query's reasoning steps. It carries the argument because it turns "correct the SQL" into "correct the trajectory", a task where schema errors and logic errors can be isolated in separate stages. In the inference loop, the Base Action Model (BAM) performs the SQL-to-trajectory conversion, the Schema Augmentation Model (SAM) masks and refills schema tokens to repair schema linking, and the Logic Optimization Model (LOM) edits the trajectory to repair logic, after which the generator LLM translates the final trajectory back to SQL. Training relies on three scripted perturbation types (ADD, DELETE, SUBSTITUTE) applied to error-free trajectories to synthesize erroneous ones, plus a hierarchical self-evolution loop where BAM, after teacher distillation, generates the augmentation data for the other two models.

What would settle it

Take the set of BIRD and SPIDER instances where SHARE still fails after correction, translate the original SQL and the final SQL into action trajectories, and measure round-trip fidelity: whether converting each trajectory back to SQL reproduces the same execution results over the database. If a substantial fraction of failures coincides with trajectories that lose or distort SQL semantics (for example, dropping join conditions or set operations), the central assumption fails; if failures occur despite faithful trajectories, the bottleneck lies elsewhere, such as in the generator's ability to read the repaired trajectory.

Watch

Extended reading notes

Core claim

The paper claims that the reason LLMs fail at self-correcting SQL is that they cannot see the reasoning behind a declarative query, and that this failure can be circumvented by an intermediate representation. SHARE converts the initial SQL into an action trajectory, a sequence of pandas-like functions such as where, select, groupby, and orderby, that mirrors the query's stepwise logic. A Schema Augmentation Model then masks and re-fills schema elements to fix schema-linking mistakes, and a Logic Optimization Model revises the trajectory to fix logical errors, with both models trained under a hierarchical self-evolution strategy in which the Base Action Model generates augmented data after initial distillation from GPT-4o. The refined trajectory is returned to the LLM, which regenerates the SQL. The paper reports that this single-round zero-shot feedback loop improves execution accuracy for several generator models, holds up with as little as 50% of the training data, and generalizes to MySQL and PostgreSQL without dialect-specific supervision.

Load-bearing premise

The pipeline assumes that the pandas-like action trajectory is a faithful, verifiable intermediate that preserves all correction-relevant meaning of the SQL query, and that the LLM can reliably convert a repaired trajectory back into correct SQL, even though neither of those steps is measured in the paper.

Editorial extensions

If this is right

  • Single-round correction with SHARE-8B outperforms the strongest compared baseline (MAGIC) by 4.61 execution-accuracy points on BIRD while costing about a tenth as much per thousand instances.
  • The same trained assistant improves accuracy for other generators, including Claude-3.5-Sonnet (up 28.64% relatively), GPT-4o-mini, Llama-3.1-70B, and a 6.7B open-source model, not just the GPT-4o teacher it was distilled from.
  • Training with only 50% of the data still beats the strongest baseline, suggesting the method is data-efficient; the paper also reports a threshold effect where gains on hard instances stay erratic until training data exceeds about 20% of the full set.
  • Without additional training, SHARE transfers to MySQL and PostgreSQL, supporting the claim that the models learn low-level reasoning-path corrections rather than dialect-specific patterns.
  • Replacing the GPT-4o teacher with Llama-3.1-70B keeps most of the gain, indicating the pipeline does not depend on the teacher's particular error patterns.

Reading between the lines

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

  • The paper never directly measures how faithfully an action trajectory preserves SQL semantics; a trajectory can be internally consistent yet drop a join condition or set-operation nuance. A direct round-trip fidelity test (trajectory to SQL back to trajectory on a held-out error set) would show whether correction failures cluster where the trajectory distorts the original query.
  • Because the assistant is demonstrated only for a single correction round, a natural next step is to test whether repeated rounds compound gains or accumulate drift; the current evidence does not rule out either outcome.
  • The trajectory abstraction is a general idea for code correction beyond SQL, a direction the paper leaves open; the mechanism only requires a set of elementary operations into which a declarative program can be decomposed.
  • The error-category analysis shows weak gains on "mathematical delusion" errors (only a 1.63% reduction), which hints that trajectory repair fixes structure and schema but does not add numerical reasoning power the generator itself lacks.
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

4 major / 6 minor

Summary. The paper proposes SHARE, a pipeline of three small language models (BAM, SAM, and LOM) that assists an LLM generator in self-correcting text-to-SQL outputs. BAM converts an initial SQL query into a pandas-like action trajectory intended to expose the query's reasoning; SAM corrects schema-linking errors in the trajectory; LOM corrects logic errors; and the generator LLM then converts the refined trajectory back into SQL. The models are trained with a GPT-4o teacher for BAM, while BAM itself synthesizes training data for SAM and LOM through masking and error perturbation. Experiments on BIRD, SPIDER, DK, and REALISTIC report execution accuracy (EX) gains over GPT-4o and several other generator models, together with ablations, a cost analysis, a low-resource data analysis, dialect-generalization experiments, and an open-source-teacher variant.

Significance. If the reported results hold, SHARE is a practical contribution: it shows that an 8B-parameter assistant can improve LLM self-correction in text-to-SQL while reducing token cost relative to calling a large proprietary model repeatedly. The paper is empirically broad, covering four benchmarks, multiple generator models, ablations that support the role of each module, a standalone-inference experiment (Appendix C.3), an overcorrection analysis (Appendix C.4), and a comparison of teacher models (Table 7). The work also ships a reproducibility commitment (code and models). However, the central mechanism depends on the faithfulness of an action-trajectory intermediate that is never directly measured on the evaluation sets, and several reporting gaps make some comparisons hard to assess.

major comments (4)
  1. [Section 3.2 and Figure 13] The paper never directly measures the fidelity of the action-trajectory intermediate on the evaluation sets. The entire mechanism assumes that BAM's trajectory faithfully encodes the SQL semantics and that the generator can reliably convert a refined trajectory back into correct SQL. The round-trip filter used during BAM data construction tests reversibility only on the training distribution, and the case study in Appendix C.6 gives one positive example. The paper should report, on the BIRD and SPIDER dev sets, (i) what fraction of BAM-generated trajectories round-trip to the initial SQL, and (ii) how often the generator's conversion of the refined trajectory produces executable SQL that matches the intended correction. Without these measurements, part of the observed gains could come from the generator's own regeneration rather than from the trajectory-level corrections made by SAM and LOM.
  2. [Section 4.2, Table 2] The key comparison table is incomplete and lacks variance information. The MAGIC row reports SPIDER only as a total (85.66) with all difficulty-level sub-scores missing, while every other method in the table has a full breakdown; no explanation is provided. In addition, all EX numbers in Tables 2–5 are point estimates, with Appendix B.2 only stating that results are 'the average of five repeated trials.' Several differences are small in absolute terms (for example, +0.24 over MAGIC on the SPIDER total and many per-difficulty gaps under 3 points), so without standard deviations, confidence intervals, or significance tests, the reader cannot tell whether those differences are meaningful. The authors should report variance or CIs and either complete or explicitly justify the MAGIC row.
  3. [Section 4.5, Figure 4] The dialect-generalization experiment is missing its experimental protocol. The text claims that SHARE performs well on MySQL and PostgreSQL 'even without additional training,' but it does not describe how the SQLite-based BIRD dev set was converted to these dialects, how the GPT-4o baseline was prompted in the new dialects, or whether the same action-trajectory inference and evaluation pipeline was used. Without this protocol, the comparison in Figure 4 is not interpretable. A concise description of the SQL translation, generation, and execution steps is needed.
  4. [Section 3.1 and Appendix A.3] The composition of LOM training data is not reported, which matters for the generalizability claim. The paper states that erroneous trajectories come from two sources: erroneous initial SQLs (generated by GPT-4o using the same baseline prompt used at evaluation) and synthetic perturbations of verified trajectories. Since the main evaluation uses GPT-4o as the generator, a large share of real GPT-4o errors in LOM training could align the assistant with GPT-4o's specific error distribution. The cross-generator results in Table 4 mitigate this concern, but the paper should state the fraction of the 15k LOM examples from each source and ideally show a version trained only on synthetic perturbations.
minor comments (6)
  1. [Section 4.7] The text says 'Table 5 presents the correction performance across fine-grained error categories,' but Table 5 is the ablation study and the fine-grained error analysis appears in Figure 5. Please correct this cross-reference.
  2. [Appendix B.2 and Section 4.2] The number of repeated trials is inconsistent: Appendix B.2 says results are the average of five repeated trials, while the low-resource paragraph says each experiment was repeated three times. Please state the number of repetitions used for each reported table.
  3. [Table 8 caption] The caption contains a typo: 'Logic Optimation Model' should be 'Logic Optimization Model.'
  4. [Section 3.4 and Algorithm 1] The value of K, the number of error perturbations per verified trajectory, is never specified. Please report the value used to construct the 15k LOM training set.
  5. [Figure 3] The x-axis labels in the rendered figure are illegible, and the text refers to 0%, 10%, 20%, 50%, and presumably 100% of training data, but these points are not clearly marked. Please redraw the figure with explicit axis labels.
  6. [Section 2.1, Eq. (2)] The symbol s_i is used for both the initial SQL and the refined SQL in the problem definition; using a prime or a different subscript for the refined query would avoid ambiguity.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: SHARE's gains are measured against held-out external benchmarks; the pandas-like action design is inherited from a self-citation but is validated independently.

full rationale

The paper's derivation chain is empirical rather than definitional. BAM is trained on 13k (SQL, trajectory) pairs filtered by a round-trip check that the trajectory can be reverted to the source SQL, which directly tests representational sufficiency; SAM and LOM are trained on schema-masked and perturbed trajectories and evaluated by downstream EX on BIRD/SPIDER/DK/REALISTIC dev sets. The main claim (14.80% and 11.41% relative EX gains with SHARE-8B) is an external benchmark result, not a quantity defined by the training objective. The only notable self-citation, Qu et al. 2024, supplies the pandas-like action vocabulary and error taxonomy; the present paper independently validates that choice through round-trip filtering, ablations (Table 5), cross-generator (Table 4), cross-dialect (Fig. 4), and low-resource (Fig. 3) experiments, so the citation is not load-bearing. The LOM training set includes GPT-4o's erroneous SQLs generated with the same baseline prompt used at evaluation, but this aligns the assistant with the generator's error distribution rather than forcing the reported outcome; cross-generator results show gains for Claude, Llama, and Qwen too. No equation in the paper reduces a predicted quantity to a fitted parameter or to the action-space definition.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claims rest on two categories of assumptions: the action-trajectory representation is faithful and reversible, and the synthetic error perturbations plus the use of GPT-4o's own errors for training are representative of the test-time error distribution. No numerical constants are fitted to benchmark results; the hand-set choices are the LOM negative-sample ratio, the unspecified K, and the strict reversibility filter.

free parameters (3)
  • LOM negative sample ratio = 4:1 (verified, verified) to (erroneous, verified)
    Appendix C.4 introduces the ratio to reduce overcorrection; it is chosen without a sweep and affects correction behavior.
  • K: number of error perturbations per verified trajectory = not stated
    Algorithm 1 uses K as an input but the paper never gives its value; it controls the size of LOM training data.
  • Reversibility filter strictness = 100% (all kept pairs must revert exactly)
    Section 3.2 keeps only (s,t) pairs where t reverts to s; the threshold is not varied or reported.
assumptions (4)
  • domain assumption Action trajectories (pandas-like stepwise operations) faithfully represent the reasoning of SQL generation and are reversible to SQL.
    Section 3.2 and Figure 2: this is the basis for BAM and for feeding trajectories to the LLM as correction feedback; no direct validation of trajectory quality other than the reversibility filter during training.
  • domain assumption The generator LLM can convert a refined trajectory back into an accurate SQL query when prompted.
    Section 3.5 and Figure 13: the final SQL is produced by the LLM from SHARE's trajectory; errors in this translation would be attributed to the assistant anyway, but the method's success depends on this step.
  • domain assumption The error distribution of the generator at test time (same baseline prompt on BIRD/SPIDER) matches the errors seen in the training set.
    Section 3.1: initial SQLs in training are from GPT-4o using the BIRD baseline prompt; the paper tests other generators, but the assistant is most aligned with GPT-4o errors.
  • ad hoc to paper The three synthetic error perturbation types (ADD, DELETE, SUBSTITUTE) are representative of real text-to-SQL errors.
    Section 3.4 and Table 1: these are the only ways errors are simulated for LOM training; if real errors differ, LOM may not correct them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SHARE: An SLM-based Hierarchical Action CorREction Assistant for Text-to-SQL." pith.science (2026). https://pith.science/paper/7WO7NIIE

@misc{pith2026250600391,
  author       = {Pith},
  title        = {Pith review of: SHARE: An SLM-based Hierarchical Action CorREction Assistant for Text-to-SQL},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7WO7NIIE}},
  note         = {Machine review of arXiv:2506.00391}
}
read the original abstract

Current self-correction approaches in text-to-SQL face two critical limitations: 1) Conventional self-correction methods rely on recursive self-calls of LLMs, resulting in multiplicative computational overhead, and 2) LLMs struggle to implement effective error detection and correction for declarative SQL queries, as they fail to demonstrate the underlying reasoning path. In this work, we propose SHARE, an SLM-based Hierarchical Action corREction assistant that enables LLMs to perform more precise error localization and efficient correction. SHARE orchestrates three specialized Small Language Models (SLMs) in a sequential pipeline, where it first transforms declarative SQL queries into stepwise action trajectories that reveal underlying reasoning, followed by a two-phase granular refinement. We further propose a novel hierarchical self-evolution strategy for data-efficient training. Experimental results demonstrate that SHARE effectively enhances self-correction capabilities while proving robust across various LLMs. Furthermore, our comprehensive analysis shows that SHARE maintains strong performance even in low-resource training settings, which is particularly valuable for text-to-SQL applications with data privacy constraints.

Figures

Figures reproduced from arXiv: 2506.00391 by the authors.

Figure 1
Figure 1. Illustrations of self-debugging and self [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An illustration of SHARE. Figure (A)-(C) illustrate the training architecture of three specialized SLMs in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The effect of training data scale on SHARE. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The performance of SHARE on BIRD across various SQL dialects, specifically MySQL (left) and PostgreSQL (right). small and large LMs. This paradigm effectively reduces the inference overhead of SHARE and incurs only one-tenth the cost of the most econom￾ical baseline. N…
Figure 5
Figure 5. Figure 5: The correction performance across fine￾grained error categories on BIRD. 4.7 Quantitative Analysis of SQL Error Corrections To quantify the effectiveness of SHARE in error correction, we analyze the SQL queries generated by the GPT-4o baseline. Following TA-SQL (Qu et …
Figure 6
Figure 6. Figure 6: Independent inference performance of SHARE on BIRD. SLM’s text-to-SQL performance without compro￾mising its existing strengths. Moreover, this effi￾cient design supports secure on-device deployment that does not require exposing sensitive database content, relying sole…
Figure 7
Figure 7. Figure 7: The prompt for the data construction process of BAM. [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]
Figure 8
Figure 8. Figure 8: The prompt for the error perturbation strategy implemented by BAM. [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: The prompt for BAM to convert SQL to action trajectory. pert. Action Trajectory (AT) is a piece of stepwise actionbased code to show the u Actions utilized in the trajectory are pandas-like functions [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 11
Figure 11. Figure 11: The prompt for SAM to reinsert the correct schema in the schema-based variant. [PITH_FULL_IMAGE:figures/full_fig_p023_11.png]
Figure 13
Figure 13. Figure 13: The prompt for LLM to generate refined SQL given the trajectory outputted by SHARE. [PITH_FULL_IMAGE:figures/full_fig_p025_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 49 canonical work pages

  1. [1]

    Anthropic. 2024. https://www.anthropic.com/news/claude-3-haiku Claude 3 haiku: our fastest model yet

    work page 2024

  2. [2]

    Arian Askari, Christian Poelitz, and Xinye Tang. 2024. http://arxiv.org/abs/2406.12692 Magic: Generating self-correction guideline for in-context text-to-sql

    arXiv 2024

  3. [3]

    Kamran Ahmad Awan, Ikram Ud Din, Ahmad Almogren, and Joel J. P. C. Rodrigues. 2023. Privacy-preserving big data security for iot with federated learning and cryptography. IEEE Access

    work page 2023

  4. [4]

    a is b" fail to learn

    Lukas Berglund, Meg Tong, Maximilian Kaufmann, Mikita Balesni, Asa Cooper Stickland, Tomasz Korbak, and Owain Evans. 2024. The reversal curse: Llms trained on "a is b" fail to learn "b is a". In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    work page 2024

  5. [5]

    Bei Chen, Fengji Zhang, Anh Nguyen, Daoguang Zan, Zeqi Lin, Jian - Guang Lou, and Weizhu Chen. 2023. Codet: Code generation with generated tests. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023

    work page 2023

  6. [6]

    Xinyun Chen, Maxwell Lin, Nathanael Sch \" a rli, and Denny Zhou. 2024 a . Teaching large language models to self-debug. In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    work page 2024

  7. [7]

    Xinyun Chen, Maxwell Lin, Nathanael Sch \" a rli, and Denny Zhou. 2024 b . Teaching large language models to self-debug. In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    work page 2024

  8. [8]

    Xiang Deng, Yu Gu, Boyuan Zheng, Shijie Chen, Sam Stevens, Boshi Wang, Huan Sun, and Yu Su. 2024. Mind2web: Towards a generalist agent for the web. Advances in Neural Information Processing Systems

    work page 2024

Show all 59 references
  1. [9]

    Longxu Dou, Yan Gao, Mingyang Pan, Dingzirui Wang, Wanxiang Che, Jian - Guang Lou, and Dechen Zhan. 2023. Unisar: a unified structure-aware autoregressive language model for text-to-sql semantic parsing. Int. J. Mach. Learn. Cybern

    2023

  2. [10]

    Nouha Dziri, Andrea Madotto, Osmar Za \" ane, and Avishek Joey Bose. 2021. Neural path hunter: Reducing hallucination in dialogue systems via path grounding. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021, Virtual Event / ...

    2021

  3. [11]

    Yujian Gan, Xinyun Chen, and Matthew Purver. 2021. Exploring underexplored limitations of cross-domain text-to-sql generalization. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021, Virtual Event / Punta Cana, Dominican Repub...

    2021

  4. [12]

    Dawei Gao, Haibin Wang, Yaliang Li, Xiuyu Sun, Yichen Qian, Bolin Ding, and Jingren Zhou. 2024. Text-to-sql empowered by large language models: A benchmark evaluation. Proc. VLDB Endow

    2024

  5. [13]

    Zhibin Gou, Zhihong Shao, Yeyun Gong, Yelong Shen, Yujiu Yang, Nan Duan, and Weizhu Chen. 2024. CRITIC: large language models can self-correct with tool-interactive critiquing. In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May...

    2024

  6. [14]

    Qiuhan Gu. 2023. Llm-based code generation method for golang compiler testing. In Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, ESEC/FSE 2023, San Francisco, CA, USA, December 3-9, 2023

    2023

  7. [15]

    Yu Gu, Yiheng Shu, Hao Yu, Xiao Liu, Yuxiao Dong, Jie Tang, Jayanth Srinivasa, Hugo Latapie, and Yu Su. 2024. Middleware for llms: Tools are instrumental for language agents in complex environments. arXiv preprint arXiv:2402.14672

    2024 arXiv

  8. [16]

    Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen - Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

    Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen - Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. Lora: Low-rank adaptation of large language models. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022

    2022

  9. [17]

    Jie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou. 2024. Large language models cannot self-correct reasoning yet. In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    2024

  10. [18]

    Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Yejin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM Comput. Surv

    2023

  11. [19]

    Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei

    Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B. Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei. 2020. Scaling laws for neural language models

    2020

  12. [20]

    Dongjun Lee, Choongwon Park, Jaehyuk Kim, and Heesoo Park. 2024. http://arxiv.org/abs/2405.07467 Mcs-sql: Leveraging multiple prompts and multiple-choice selection for text-to-sql generation

    2024 arXiv

  13. [21]

    Wenqiang Lei, Weixin Wang, Zhixin Ma, Tian Gan, Wei Lu, Min-Yen Kan, and Tat-Seng Chua. 2020. Re-examining the role of schema linking in text-to- SQL . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)

    2020

  14. [22]

    Jinyang Li, Binyuan Hui, Reynold Cheng, Bowen Qin, Chenhao Ma, Nan Huo, Fei Huang, Wenyu Du, Luo Si, and Yongbin Li. 2023. Graphix-t5: Mixing pre-trained transformers with graph-aware layers for text-to-sql parsing. In Thirty-Seventh AAAI Conference on Artificial Intelligence,...

    2023

  15. [23]

    Jinyang Li, Binyuan Hui, Ge Qu, Jiaxi Yang, Binhua Li, Bowen Li, Bailin Wang, Bowen Qin, Ruiying Geng, Nan Huo, et al. 2024 a . Can llm already serve as a database interface? a big bench for large-scale database grounded text-to-sqls. Advances in Neural Information Processing Systems

    2024

  16. [24]

    Jinyang Li, Nan Huo, Yan Gao, Jiayi Shi, Yingxiu Zhao, Ge Qu, Yurong Wu, Chenhao Ma, Jian-Guang Lou, and Reynold Cheng. 2024 b . Tapilot-crossing: Benchmarking and evolving llms towards interactive data analysis agents

    2024

  17. [25]

    Zhenwen Li and Tao Xie. 2024. Using llm to select the right sql query from candidates

    2024

  18. [26]

    Huihui Liu, Yiding Yang, and Xinchao Wang. 2021. Overcoming catastrophic forgetting in graph neural networks. In Thirty-Fifth AAAI Conference on Artificial Intelligence, AAAI 2021, Thirty-Third Conference on Innovative Applications of Artificial Intelligence, IAAI 2021, The El...

    2021

  19. [27]

    Xiping Liu and Zhao Tan. 2024. http://arxiv.org/abs/2404.14453 Epi-sql: Enhancing text-to-sql translation with error-prevention instructions

    2024 arXiv

  20. [28]

    Seyed - Iman Mirzadeh, Keivan Alizadeh, Hooman Shahrokhi, Oncel Tuzel, Samy Bengio, and Mehrdad Farajtabar. 2025. Gsm-symbolic: Understanding the limitations of mathematical reasoning in large language models. In The Thirteenth International Conference on Learning Representati...

    2025

  21. [29]

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F. Christiano, Jan Leik...

    2022

  22. [30]

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F. Christiano, Jan Leik...

    2022

  23. [31]

    Liangming Pan, Michael Saxon, Wenda Xu, Deepak Nathani, Xinyi Wang, and William Yang Wang. 2024 a . Automatically correcting large language models: Surveying the Landscape of Diverse Automated Correction Strategies. Trans. Assoc. Comput. Linguistics

    2024

  24. [32]

    Liangming Pan, Michael Saxon, Wenda Xu, Deepak Nathani, Xinyi Wang, and William Yang Wang. 2024 b . Automatically correcting large language models: Surveying the landscape of diverse automated correction strategies. Transactions of the Association for Computational Linguistics

    2024

  25. [33]

    Tianyu Peng and Jiajun Zhang. 2024. http://arxiv.org/abs/2409.12545 Enhancing knowledge distillation of large language models through efficient multi-modal distribution alignment

    2024 arXiv

  26. [34]

    Mohammadreza Pourreza, Hailong Li, Ruoxi Sun, Yeounoh Chung, Shayan Talaei, Gaurav Tarlok Kakkar, Yu Gan, Amin Saberi, Fatma Ozcan, and Sercan O. Arik. 2024. Chase-sql: Multi-path reasoning and preference optimized candidate selection in text-to-sql

    2024

  27. [35]

    Mohammadreza Pourreza and Davood Rafiei. 2024. Din-sql: Decomposed in-context learning of text-to-sql with self-correction. Advances in Neural Information Processing Systems

    2024

  28. [36]

    Bowen Qin, Lihan Wang, Binyuan Hui, Bowen Li, Xiangpeng Wei, Binhua Li, Fei Huang, Luo Si, Min Yang, and Yongbin Li. 2022. SUN: exploring intrinsic uncertainties in text-to-sql parsers. In Proceedings of the 29th International Conference on Computational Linguistics, COLING 20...

    2022

  29. [37]

    XiPeng Qiu, TianXiang Sun, YiGe Xu, YunFan Shao, Ning Dai, and XuanJing Huang. 2020. Pre-trained models for natural language processing: A survey. Science China Technological Sciences

    2020

  30. [38]

    Ge Qu, Jinyang Li, Bowen Li, Bowen Qin, Nan Huo, Chenhao Ma, and Reynold Cheng. 2024. Before generation, align it! A novel and effective strategy for mitigating hallucinations in text-to-sql generation. Association for Computational Linguistics

    2024

  31. [39]

    Manning, Stefano Ermon, and Chelsea Finn

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D. Manning, Stefano Ermon, and Chelsea Finn. 2023. Direct preference optimization: Your language model is secretly a reward model. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural In...

    2023

  32. [40]

    Nitarshan Rajkumar, Raymond Li, and Dzmitry Bahdanau. 2022. Evaluating the text-to-sql capabilities of large language models

    2022

  33. [41]

    Sithursan Sivasubramaniam, Cedric Osei-Akoto, Yi Zhang, Kurt Stockinger, and Jonathan Fuerst. 2024. Sm3-text-to-query: Synthetic multi-model medical text-to-query benchmark

    2024

  34. [42]

    Shayan Talaei, Mohammadreza Pourreza, Yu-Chen Chang, Azalia Mirhoseini, and Amin Saberi. 2024. Chess: Contextual harnessing for efficient sql synthesis

    2024

  35. [43]

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, and Andrew M. 2024. Gemini: A family of highly capable multimodal models

    2024

  36. [44]

    Bing Wang, Changyu Ren, Jian Yang, Xinnian Liang, Jiaqi Bai, Linzheng Chai, Zhao Yan, Qian-Wen Zhang, Di Yin, Xing Sun, and Zhoujun Li. 2024. http://arxiv.org/abs/2312.11242 Mac-sql: A multi-agent collaborative framework for text-to-sql

    2024 arXiv

  37. [45]

    Le, Ed H

    Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc V. Le, Ed H. Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2023. Self-consistency improves chain of thought reasoning in language models. In The Eleventh International Conference on Learning Representations, ICLR 2023, K...

    2023

  38. [46]

    this is my sql, are you with me?

    Hanchen Xia, Feng Jiang, Naihao Deng, Cunxiang Wang, Guojiang Zhao, Rada Mihalcea, and Yue Zhang. 2024. r^3 : "this is my sql, are you with me?" a consensus-based multi-agent system for text-to-sql tasks

    2024

  39. [47]

    Yuanzhen Xie, Xinzhou Jin, Tao Xie, Matrixmxlin Matrixmxlin, Liang Chen, Chenyun Yu, Cheng Lei, Chengxiang Zhuo, Bo Hu, and Zang Li. 2024. Decomposition for enhancing attention: Improving llm-based text-to-sql through workflow paradigm. In Findings of the Association for Compu...

    2024

  40. [48]

    Shusheng Xu, Wei Fu, Jiaxuan Gao, Wenjie Ye, Weilin Liu, Zhiyu Mei, Guangju Wang, Chao Yu, and Yi Wu. 2024 a . Is dpo superior to ppo for llm alignment? a comprehensive study. In Proceedings of the 41st International Conference on Machine Learning

    2024

  41. [49]

    Xiaohan Xu, Ming Li, Chongyang Tao, Tao Shen, Reynold Cheng, Jinyang Li, Can Xu, Dacheng Tao, and Tianyi Zhou. 2024 b . A survey on knowledge distillation of large language models

    2024

  42. [50]

    Pengcheng Yin and Graham Neubig. 2017. A syntactic neural model for general-purpose code generation. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, ACL 2017, Vancouver, Canada, July 30 - August 4, Volume 1: Long Papers

    2017

  43. [51]

    Tao Yu, Rui Zhang, Kai Yang, Michihiro Yasunaga, Dongxu Wang, Zifan Li, James Ma, Irene Li, Qingning Yao, Shanelle Roman, Zilin Zhang, and Dragomir R. Radev. 2018. Spider: A large-scale human-labeled dataset for complex and cross-domain semantic parsing and text-to-sql task. I...

    2018

  44. [52]

    Jianguo Zhang, Tian Lan, Ming Zhu, Zuxin Liu, Thai Hoang, Shirley Kokane, Weiran Yao, Juntao Tan, Akshara Prabhakar, Haolin Chen, Zhiwei Liu, Yihao Feng, Tulika Awalgaonkar, Rithesh Murthy, Eric Hu, Zeyuan Chen, Ran Xu, Juan Carlos Niebles, Shelby Heinecke, Huan Wang, Silvio S...

    2024 arXiv

  45. [53]

    Xuanliang Zhang, Dingzirui Wang, Longxu Dou, Qingfu Zhu, and Wanxiang Che. 2025. A survey of table reasoning with large language models. Frontiers Comput. Sci

    2025

  46. [54]

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, Yifan Du, Chen Yang, Yushuo Chen, Zhipeng Chen, Jinhao Jiang, Ruiyang Ren, Yifan Li, Xinyu Tang, Zikang Liu, Peiyu Liu, Jian-Yun Nie, and Ji-Rong W...

    2024

  47. [55]

    Xuanle Zhao, Xianzhen Luo, Qi Shi, Chi Chen, Shuo Wang, Wanxiang Che, Zhiyuan Liu, and Maosong Sun. 2025. Chartcoder: Advancing multimodal large language model for chart-to-code generation

    2025

  48. [56]

    Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, and Zheyan Luo. 2024. L lama F actory: Unified efficient fine-tuning of 100+ language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations). Ass...

    2024

  49. [57]

    Ruiqi Zhong, Charlie Snell, Dan Klein, and Jason Eisner. 2023. Non-programmers can label programs indirectly via active examples: A case study with text-to-sql. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, Singapore, De...

    2023

  50. [58]

    URL: " 'urlintro :=

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

  51. [59]

    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.