Pith. sign in

REVIEW 2 major objections 7 minor 31 references

Extract Information from Hybrid Long Documents Leveraging LLMs: A Framework and Dataset

T0 review · 2 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read LLMs can pull numbers from hybrid long documents by retrieving, summarizing, and extracting only the relevant segments.

desk verdict Systematic ablations and a new financial dataset, but the undefined baseline and test-set tuning make the headline numbers provisional; worth reviewing with major revisions. read the letter →

arxiv 2412.20072 v2 pith:NOEH3LWE submitted 2024-12-28 cs.CL cs.AI

Chongjian Yue , Xinrun Xu , Xiaojun Ma , Lun Du , Zhiming Ding , Shi Han , Dongmei Zhang , Qi Zhang This is my paper · ORCID
classification cs.CLcs.AI
keywords hybridlongdocumentsinformationextractionlargelanguagemodelstableserializationembeddingretrievalfinancialreportsfew-shotpromptingRETAmetric
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

Large language models struggle with hybrid long documents—reports where prose and tables are woven together and the whole text exceeds the model's context window. This paper proposes AIE, a four-stage pipeline that splits the document into segments, retrieves the segments most similar to a query keyword using embedding similarity, condenses them with an LLM, and extracts the target number. On the new FINE financial dataset with GPT-3.5, AIE reaches 63.89% accuracy at a 1% relative-error tolerance, versus 13.92% for naive truncation, and it also beats the naive approach on Wikipedia and materials-science documents. The paper further reports that plain-text table serialization outperforms hierarchical XML and HTML, that retrieving three segments is the best setting, and that one well-chosen few-shot example beats zero or multiple shots.

What carries the argument

The central object is the AIE pipeline with four modules: Segmentation, Retrieval, Summarization, and Extraction. Segmentation serializes tables into plain text, splits oversized elements, and merges adjacent small ones; Retrieval scores each segment against the query keyword using a sentence-embedding model (SentenceTransformer) and keeps the top K=3 segments; Summarization applies the Refine strategy, iteratively updating a running summary with each retrieved segment; Extraction uses a tailored prompt to return the precise value. The evaluation machinery is the RETA metric, which counts a prediction as correct when its relative error stays within a chosen tolerance, and the new FINE dataset of financial KPI values extracted from SEC filings.

What would settle it

Build a hybrid document in which the target value sits in a table with paraphrased headers and no exact keyword match, then check whether embedding retrieval selects the segment containing the value; if retrieval recall on such tables drops sharply, the central claim fails at its first step. A second check is to evaluate the R@n choice on a held-out split rather than the analyzed test set, since the R@3 advantage could shrink or reverse on a new split.

Watch

Extended reading notes

Core claim

The paper's central claim is that information extraction from hybrid long documents does not require processing the entire document or preserving the hierarchical structure of tables. A pipeline that segments the document, retrieves keyword-similar segments by embedding similarity, summarizes them iteratively with an LLM, and extracts the answer from the summary consistently outperforms a naive method that truncates the document, across financial reports, Wikipedia pages, and materials-science papers. On FINE, AIE with GPT-3.5 achieves an average accuracy of 69.93% versus 15.99% for the naive method. The paper also claims that a simple serialization of tables into plain text is sufficient for LLM table comprehension, and that the framework benefits from retrieving three segments, using the Refine summarization strategy, completing query keywords with company and time context, and using a single precision-aware example.

Load-bearing premise

The retrieval step trusts that embedding similarity between the query keyword and a document segment points to the segment containing the answer; if the keyword is phrased differently from the table or the table serializes poorly, retrieval can select the wrong segments and the later summarization and extraction steps cannot recover the value.

Editorial extensions

If this is right

  • A retrieval-first pipeline removes the context-window ceiling: LLMs can work on arbitrarily long hybrid documents by reading only the segments that match the query.
  • Plain-text table serialization is enough for LLM table understanding, so expensive structure-preserving encodings such as XML or HTML are unnecessary and can even hurt accuracy.
  • The number of retrieved segments has a sweet spot: R@3 outperforms R@1, R@5, and R@7 on FINE, so retrieval quantity should be tuned rather than maximized.
  • Keyword completion with document context (company and time) materially improves extraction, raising average accuracy from 0.3922 with the bare keyword to 0.6993 with full context.
  • A single well-designed few-shot example gives better accuracy than zero-shot or multiple shots, indicating that prompt consistency matters more than example count.

Reading between the lines

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

  • If embedding retrieval is the bottleneck, then better embeddings for table-heavy text, or hybrid retrieval that also uses table structure, should directly raise extraction accuracy within the same pipeline.
  • The R@3 optimum found on FINE may not transfer to other genres; a production version could select the number of retrieved segments per query based on retrieval confidence rather than a global constant.
  • The framework's modularity suggests it could be applied to legal, medical, or scientific records, but the paper demonstrates only finance, Wikipedia, and materials-science domains, so cross-domain generality remains an open extension.
  • The plain-serialization result hints that LLMs may understand tables through token co-occurrence between headers and values rather than through structural nesting, which could be probed by shuffling table layout while keeping content identical.
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

2 major / 7 minor

Summary. The paper proposes AIE, a four-module framework (Segmentation, Retrieval, Summarization, Extraction) for enabling LLMs to perform information extraction from hybrid long documents (HLDs) that exceed the LLM token limit. It introduces FINE, a new dataset of financial key performance indicators extracted from SEC filings, and evaluates AIE against a 'naive' LLM-based approach on FINE, WIKIR, and MPP. The experiments show large improvements for AIE on FINE (63.89% vs 13.92% at RETA 1% with GPT-3.5) and analyze several design choices: table serialization formats, number of retrieved segments, summarization strategies, numerical precision prompts, keyword completion, and few-shot counts. The paper concludes that a retrieve-summarize-extract pipeline with plain-table serialization and refine summarization is effective for HLD information extraction.

Significance. If the results are reproducible, the AIE framework offers a practical recipe for LLM-based extraction from long, mixed text-table documents, and the FINE dataset is a valuable new resource for the community. The paper also provides a useful systematic comparison of multiple design decisions, including table serialization, retrieval granularity, and prompt engineering. The main claims are concrete and testable, and the dataset/code release is a positive feature. However, the significance is currently tempered by two methodological issues: an undefined baseline and test-set-driven configuration selection, both of which affect the credibility of the quantitative claims.

major comments (2)
  1. [Section IV, Figures 2–4] The 'naive LLM-based approach' is never defined. Please specify exactly how the naive method processes an HLD: whether the document is truncated, how tables are serialized, which prompt is used for extraction, whether the same extraction module is used, and how the final answer is produced. Without this definition, the comparison in Figures 2–4 and the central claim that 'AIE consistently outperforms the naive method' are not interpretable or reproducible.
  2. [Section IV, Tables II–VII] The final AIE configuration is selected by running all variants on the FINE test set and picking the winner in each dimension (R@3 in Table III, PLAIN in Table II, Refine in Table IV, TD-RSP in Table V, KTC in Table VI, 1-shot in Table VII). The headline AIE accuracy in Figure 2 is then reported under this configuration. This is test-set selection, so the reported numbers are upper bounds rather than unbiased estimates of performance. Please enforce a validation split (or cross-validation) for model selection and report the chosen configuration's accuracy on a held-out test set, or at least discuss the magnitude of the selection effect.
minor comments (7)
  1. [Abstract] The phrase 'The naive AIE has adaptability in many complex scenarios' appears to be a typo; it should likely read 'The AIE has adaptability' or 'The proposed AIE is adaptable.'
  2. [Section III] The RETA metric is described only verbally. Please provide the explicit formula and specify how predictions with missing or non-numeric outputs are scored.
  3. [Figure 2] The y-axis label 'Average Accuracy' is confusing because the x-axis already shows RETA levels. Clarify that the bars represent the average accuracy over all queries at the corresponding RETA threshold.
  4. [Figure 4] The heatmaps lack a legend and a clear description of the color scale. Add a legend and explain what the colors represent for the RPD values.
  5. [Section I and Section VI] The paper states that the dataset and code are 'publicly available in the attachments,' but no link or repository identifier is provided. Include a working URL for reproducibility.
  6. [Section IV] No variance or significance tests are reported for any of the comparisons. Given the large margins this is not fatal, but including error bars or a significance test for the headline FINE result would strengthen the claim.
  7. [Table I] For the FINE dataset, please also report the number of documents and the number of queries, not only token length statistics, so readers can gauge the dataset size.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline AIE comparison is reported under a configuration whose components (PLAIN, R@3, Refine, TD-RSP, KTC, 1-shot) were each selected as the best performer on the same FINE test set, so the central 'consistent outperformance' claim is an in-sample selection result rather than an independent prediction.

  1. fitted input called prediction [Section I, final paragraph; Section IV, Tables II-VII and Figure 2]
    "The subsequent sections detail the AIE framework, incorporating globally optimal settings determined through our analysis (Section II). ... R@3 achieves the highest accuracy across all RETA levels. ... the 1-shot setting achieves the highest accuracy across all RETA levels."

    The final AIE pipeline is assembled from the winners of a model-selection sweep on the same FINE test set: PLAIN in Table II, R@3 in Table III, Refine in Table IV, TD-RSP in Table V, KTC in Table VI, and 1-shot in Table VII. The headline accuracy (Figure 2a, 63.89% at RETA 1%) is then computed on that same test set with that selected configuration. The 'consistent outperformance' claim therefore does not test a fixed a-priori framework; it reports the performance of the best configuration on the data used to pick the configuration. The reported margins are upper bounds, and the conclusion is partly forced by the selection procedure rather than derived from an independent prediction.

full rationale

The core derivation chain (segment, retrieve by SentenceTransformer similarity, summarize, extract) is self-contained and is not definitionally tied to the target values; the FINE dataset is new and externally grounded in SEC filings, and the WIKIR/MPP evaluations provide some held-out evidence with settings fixed from FINE. However, the paper's main quantitative claim about AIE's superiority is weakened by in-sample configuration selection: every component of the reported AIE is chosen because it maximized accuracy on the same FINE test set that produces the headline numbers. This is the 'fitted input called prediction' pattern, not a definitional equivalence, so the score is partial (6) rather than total. The undefined 'naive' baseline is a reproducibility problem but not a circularity, and the self-citation [5] is not load-bearing for the derivation chain.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The framework depends on six test-set-selected design choices (K=3, PLAIN, Refine, TD-RSP, KTC, 1-shot), several domain assumptions about retrieval quality and ground truth, and no invented theoretical entities. The central empirical claims are therefore not derived from first principles; they rest on assumptions and in-sample tuning that should be validated on held-out data.

free parameters (6)
  • Retrieval segment count K = 3 (R@3)
    Selected by highest average accuracy on the FINE test set in Table III (R@3=0.6993 vs R@2=0.6743, R@5=0.6832). This is a test-set choice, not a validated hyperparameter.
  • Table serialization format = PLAIN
    PLAIN gives the highest average accuracy in Table II (0.6993 vs CSV 0.6911); the difference is small and within plausible noise.
  • Summarization strategy = Refine
    Refine beats Map-Reduce in Table IV (0.6993 vs 0.5840); the paper notes a time/accuracy tradeoff but adopts Refine.
  • Numerical precision prompt = TD-RSP
    TD-RSP yields highest accuracy in Table V (0.5748 vs TD-SP 0.5484); chosen on the same test set.
  • Keyword completion context = KTC (keyword+time+company)
    Table VI shows KTC highest; this is a deliberate design choice informed by test performance.
  • Number of shots = 1
    Table VII shows 1-shot highest (0.6993 vs 2-shot 0.6807); selected on test set.
assumptions (5)
  • domain assumption Embedding similarity between the keyword and document segments reliably locates segments containing the target value.
    The retrieval module (Section II, Retrieval Module) assumes SentenceTransformer similarity is a sufficient relevance signal for financial KPI keywords; if retrieval misses the right segment, later modules cannot recover it.
  • domain assumption Financial reports from SEC EDGAR contain a single correct numeric value for each keyword-document pair.
    FINE construction (Section III) assumes unambiguous ground truth exists in the document; ambiguous 'Revenue vs Total Net Sales' analysis suggests this is not always true, complicating the assumption.
  • domain assumption LLM extraction from a generated summary does not introduce systematic hallucination that dominates the measured errors.
    The Extraction Module relies on LLMs to output the exact value from a summary; no faithfulness calibration or error analysis is reported.
  • domain assumption The 'naive' baseline is a fair representation of direct LLM-based extraction without the AIE pipeline.
    The baseline is not described (Section IV), so the comparison assumes it is a reasonable trivial approach rather than a deliberately weak strawman.
  • standard math Cosine similarity as used in Sentence-BERT is an appropriate relevance metric.
    Uses Sentence-Transformer model [13]; standard text embedding similarity, not contested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Extract Information from Hybrid Long Documents Leveraging LLMs: A Framework and Dataset." pith.science (2026). https://pith.science/paper/NOEH3LWE

@misc{pith2026241220072,
  author       = {Pith},
  title        = {Pith review of: Extract Information from Hybrid Long Documents Leveraging LLMs: A Framework and Dataset},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NOEH3LWE}},
  note         = {Machine review of arXiv:2412.20072}
}
read the original abstract

Large Language Models (LLMs) demonstrate exceptional performance in textual understanding and tabular reasoning tasks. However, their ability to comprehend and analyze hybrid text, containing textual and tabular data, remains unexplored. The hybrid text often appears in the form of hybrid long documents (HLDs), which far exceed the token limit of LLMs. Consequently, we apply an Automated Information Extraction framework (AIE) to enable LLMs to process the HLDs and carry out experiments to analyse four important aspects of information extraction from HLDs. Given the findings: 1) The effective way to select and summarize the useful part of a HLD. 2) An easy table serialization way is enough for LLMs to understand tables. 3) The naive AIE has adaptability in many complex scenarios. 4) The useful prompt engineering to enhance LLMs on HLDs. To address the issue of dataset scarcity in HLDs and support future work, we also propose the Financial Reports Numerical Extraction (FINE) dataset. The dataset and code are publicly available in the attachments.

Figures

Figures reproduced from arXiv: 2412.20072 by the authors.

Figure 1
Figure 1. The AIE framework illustrates the end-to-end IE process, consisting of four modules: Segmentation, Retrieval, [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Comparison of the Naive method and AIE at different [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Comparison of the Naive method and AIE on WIKIR [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Exploring the Capability to Handle Keyword Ambiguity: Comparison of Naive and AIE on RPD [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Illustration of the Map-Reduce Strategy. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 13 canonical work pages

  1. [1]

    Chain-of-thought prompting elicits reasoning in large language models,

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, B. Ichter, F. Xia, E. Chi, Q. Le, and D. Zhou, “Chain-of-thought prompting elicits reasoning in large language models,” 2023

    2023

  2. [2]

    Text2analysis: A benchmark of table question answering with advanced data analysis and unclear queries,

    X. He, M. Zhou, X. Xu, X. Ma, R. Ding, L. Du, Y . Gao, R. Jia, X. Chen, S. Han et al., “Text2analysis: A benchmark of table question answering with advanced data analysis and unclear queries,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 16, 2024, pp. 18 206–18 215

    work page 2024

  3. [3]

    A survey on game playing agents and large models: Methods, applications, and challenges,

    X. Xu, Y . Wang, C. Xu, Z. Ding, J. Jiang, Z. Ding, and B. F. Karlsson, “A survey on game playing agents and large models: Methods, applications, and challenges,” arXiv preprint arXiv:2403.10249 , 2024

    arXiv 2024

  4. [4]

    Strago: Harnessing strategic guidance for prompt opti- mization,

    Y . Wu, Y . Gao, B. Zhu, Z. Zhou, X. Sun, S. Yang, J.-G. Lou, Z. Ding, and L. Yang, “Strago: Harnessing strategic guidance for prompt opti- mization,” in Findings of the Association for Computational Linguistics: EMNLP 2024, 2024, pp. 10 043–10 061

    work page 2024

  5. [5]

    Enabling and analyzing how to efficiently extract information from hybrid long documents with llms,

    C. Yue, X. Xu, X. Ma, L. Du, H. Liu, Z. Ding, Y . Jiang, S. Han, and D. Zhang, “Enabling and analyzing how to efficiently extract information from hybrid long documents with llms,” arXiv preprint arXiv:2305.16344, 2023

    arXiv 2023

  6. [6]

    Large lan- guage models are zero-shot reasoners,

    T. Kojima, S. S. Gu, M. Reid, Y . Matsuo, and Y . Iwasawa, “Large lan- guage models are zero-shot reasoners,” Advances in neural information processing systems, vol. 35, pp. 22 199–22 213, 2022

    2022

  7. [7]

    Large language models are few (1)-shot table reasoners,

    W. Chen, “Large language models are few (1)-shot table reasoners,” arXiv preprint arXiv:2210.06710 , 2022

    arXiv 2022

  8. [8]

    Large language models are versatile decomposers: Decompose evidence and questions for table-based reasoning,

    Y . Ye, B. Hui, M. Yang, B. Li, F. Huang, and Y . Li, “Large language models are versatile decomposers: Decompose evidence and questions for table-based reasoning,” arXiv preprint arXiv:2301.13808 , 2023

    arXiv 2023

Show all 31 references
  1. [9]

    Cradle: Empowering foundation agents towards general computer control,

    W. Tan, W. Zhang, X. Xu, H. Xia, Z. Ding, B. Li, B. Zhou, J. Yue, J. Jiang, Y . Li et al. , “Cradle: Empowering foundation agents towards general computer control,” in NeurIPS 2024 Workshop on Open-World Agents

    2024

  2. [10]

    Embedding-based product retrieval in taobao search,

    S. Li, F. Lv, T. Jin, G. Lin, K. Yang, X. Zeng, X.-M. Wu, and Q. Ma, “Embedding-based product retrieval in taobao search,” in Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, 2021, pp. 3181–3189

    2021

  3. [11]

    Can large lan- guage models recall reference location like humans?

    Y . Wang, X. Xu, R. Xie, W. Hu, and W. Ye, “Can large lan- guage models recall reference location like humans?” arXiv preprint arXiv:2402.17010, 2024

    2024

  4. [12]

    Mllm as retriever: Interactively learning multimodal retrieval for embodied agents,

    J. Yue, X. Xu, B. F. Karlsson, and Z. Lu, “Mllm as retriever: Interactively learning multimodal retrieval for embodied agents,” arXiv preprint arXiv:2410.03450, 2024

    2024 arXiv

  5. [13]

    Sentence-bert: Sentence embeddings using siamese bert-networks,

    N. Reimers and I. Gurevych, “Sentence-bert: Sentence embeddings using siamese bert-networks,” arXiv preprint arXiv:1908.10084 , 2019

    1908 arXiv

  6. [14]

    From dataset recycling to multi-property extraction and beyond,

    T. Dwojak, M. Pietruszka, L. Borchmann, J. Chledowski, and F. Gralin- ski, “From dataset recycling to multi-property extraction and beyond,” CoRR, vol. abs/2011.03228, 2020

    2011 arXiv

  7. [15]

    Flexible, model-agnostic method for materials data extraction from text using general purpose language models,

    M. P. Polak, S. Modi, A. Latosinska, J. Zhang, C.-W. Wang, S. Wang, A. D. Hazra, and D. Morgan, “Flexible, model-agnostic method for materials data extraction from text using general purpose language models,” arXiv preprint arXiv:2302.04914 , 2023

    2023 arXiv

  8. [16]

    A hybrid ai tool to extract key performance indicators from financial reports for benchmarking,

    E. Brito, R. Sifa, C. Bauckhage, R. Loitz, U. Lohmeier, and C. P ¨unt, “A hybrid ai tool to extract key performance indicators from financial reports for benchmarking,” in Proceedings of the ACM Symposium on Document Engineering 2019 , 2019, pp. 1–4

    2019

  9. [17]

    Spot: A tool for identifying operating segments in financial tables,

    Z. Ma, S. Pomerville, M. Di, and A. Nourbakhsh, “Spot: A tool for identifying operating segments in financial tables,” in Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval , 2020, pp. 2157–2160

    2020

  10. [18]

    Exploring word representations on time expression recognition,

    S. Chen, G. Wang, and B. F. Karlsson, “Exploring word representations on time expression recognition,” Microsoft Research Technical Report MSR-TR-2019-46, June 2019

    2019

  11. [19]

    Kpi-bert: A joint named entity recognition and rela- tion extraction model for financial reports,

    L. Hillebrand, T. Deußer, T. Dilmaghani, B. Kliem, R. Loitz, C. Bauck- hage, and R. Sifa, “Kpi-bert: A joint named entity recognition and rela- tion extraction model for financial reports,” in 2022 26th International Conference on Pattern Recognition (ICPR) . IEEE, 2022, pp. 606–612

    2022

  12. [20]

    Finqa: A dataset of numerical reasoning over financial data,

    Z. Chen, W. Chen, C. Smiley, S. Shah, I. Borova, D. Langdon, R. Moussa, M. Beane, T.-H. Huang, B. Routledge et al. , “Finqa: A dataset of numerical reasoning over financial data,” arXiv preprint arXiv:2109.00122, 2021

    2021 arXiv

  13. [21]

    Tat-qa: A question answering benchmark on a hybrid of tabular and textual content in finance,

    F. Zhu, W. Lei, Y . Huang, C. Wang, S. Zhang, J. Lv, F. Feng, and T.-S. Chua, “Tat-qa: A question answering benchmark on a hybrid of tabular and textual content in finance,” 2021

    2021

  14. [22]

    MultiHiertt: Numerical reasoning over multi hierarchical tabular and textual data,

    Y . Zhao, Y . Li, C. Li, and R. Zhang, “MultiHiertt: Numerical reasoning over multi hierarchical tabular and textual data,” in Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) . Dublin, Ireland: Association for Com...

    2022

  15. [23]

    Unleashing infinite-length input capacity for large-scale lan- guage models with self-controlled memory system,

    X. Liang, B. Wang, H. Huang, S. Wu, P. Wu, L. Lu, Z. Ma, and Z. Li, “Unleashing infinite-length input capacity for large-scale lan- guage models with self-controlled memory system,” arXiv preprint arXiv:2304.13343, 2023

    2023 arXiv

  16. [24]

    Eval- uating chatgpt’s information extraction capabilities: An assessment of performance, explainability, calibration, and faithfulness,

    B. Li, G. Fang, Y . Yang, Q. Wang, W. Ye, W. Zhao, and S. Zhang, “Eval- uating chatgpt’s information extraction capabilities: An assessment of performance, explainability, calibration, and faithfulness,”arXiv preprint arXiv:2304.11633, 2023

    2023 arXiv

  17. [25]

    Zero-shot information extraction via chatting with chatgpt,

    X. Wei, X. Cui, N. Cheng, X. Wang, X. Zhang, S. Huang, P. Xie, J. Xu, Y . Chen, M. Zhanget al., “Zero-shot information extraction via chatting with chatgpt,” arXiv preprint arXiv:2302.10205 , 2023

    2023 arXiv

  18. [26]

    Context-ner: Contextual phrase generation at scale,

    H. Gupta, S. Verma, T. Kumar, S. Mishra, T. Agrawal, A. Badugu, and H. S. Bhatt, “Context-ner: Contextual phrase generation at scale,” arXiv preprint arXiv:2109.08079, 2021

    2021 arXiv

  19. [27]

    Gpt-ner: Named entity recognition via large language models,

    S. Wang, X. Sun, X. Li, R. Ouyang, F. Wu, T. Zhang, J. Li, and G. Wang, “Gpt-ner: Named entity recognition via large language models,” arXiv preprint arXiv:2304.10428, 2023

    2023 arXiv

  20. [28]

    Gpt-re: In-context learning for relation extraction using large language models,

    Z. Wan, F. Cheng, Z. Mao, Q. Liu, H. Song, J. Li, and S. Kurohashi, “Gpt-re: In-context learning for relation extraction using large language models,” arXiv preprint arXiv:2305.02105 , 2023

    2023 arXiv

  21. [29]

    How to unleash the power of large language models for few-shot relation extraction?

    X. Xu, Y . Zhu, X. Wang, and N. Zhang, “How to unleash the power of large language models for few-shot relation extraction?” arXiv preprint arXiv:2305.01555, 2023

    2023 arXiv

  22. [30]

    Chat- graph: Interpretable text classification by converting chatgpt knowledge to graphs,

    Y . Shi, H. Ma, W. Zhong, G. Mai, X. Li, T. Liu, and J. Huang, “Chat- graph: Interpretable text classification by converting chatgpt knowledge to graphs,” arXiv preprint arXiv:2305.03513 , 2023

    2023 arXiv

  23. [31]

    Language models enable simple systems for gen- erating structured views of heterogeneous data lakes,

    S. Arora, B. Yang, S. Eyuboglu, A. Narayan, A. Hojel, I. Trum- mer, and C. R ´e, “Language models enable simple systems for gen- erating structured views of heterogeneous data lakes,” arXiv preprint arXiv:2304.09433, 2023

    2023 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.