Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Hierarchical Document Refinement for Long-context Retrieval-augmented Generation

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read LongRefiner shows that turning long documents into a section tree lets a refiner keep only query-relevant nodes, matching or beating full-document RAG with roughly one-tenth the tokens.

desk verdict A solid, useful RAG refiner whose abstract overclaims a 10x win, and whose adaptive query weight may be nearly constant as written—but the core empirical result is credible and deserves refereeing. read the letter →

arxiv 2505.10413 v1 pith:G7FHE7EO submitted 2025-05-15 cs.CL

Jiajie Jin , Xiaoxi Li , Guanting Dong , Yuyao Zhang , Yutao Zhu , Yongkang Wu , Zhonghua Li , Qi Ye
show 1 more author
Zhicheng Dou
This is my paper · ORCID
classification cs.CL
keywords retrieval-augmentedgenerationlong-contextdocumentsdocumentrefinementhierarchicaltreepromptcompressionquery-awarenodeselectionmulti-taskLoRAopen-domainquestionanswering
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

Long-context RAG is held back by retrieved documents that are far longer than the relevant parts: feeding everything to the generator is slow and costly, while chunk- or perplexity-based compression drops evidence. LongRefiner proposes turning each retrieved document into an XML-style tree of sections, subsections, and paragraphs, then selecting the tree nodes the query actually needs. The paper's claim is that this structure-aware selection, run within a 2k-token budget, outperforms existing refinement methods on all seven QA datasets tested and beats the full-content setting on six of them, while using roughly one-tenth the generator tokens and about one-quarter of the online latency. If that is right, long-context RAG becomes a cheap preprocessing problem rather than an expensive generation problem.

What carries the argument

The central object is the structured document tree $D_{\text{str}} = (N,R)$, represented in the paper's own flat XML syntax: `<section: {title}>` and `<subsection: {title}>` mark the hierarchy, `<br>` marks paragraph breaks, and `<skip>` compresses each block to its first and last $k$ tokens so the learning target stays short. This representation is what turns refinement into node selection. The selection score is $\text{Score}(n_i) = \text{LS}(n_i) + R_q \cdot \text{GS}(n_i)$: local scores are averaged from leaves to roots, global scores are split uniformly from roots to leaves, and $R_q = \text{Softmax}(P_{\text{Local}}, P_{\text{Global}})_g$ is the continuous query-scope weight. The machinery's role is to make the chosen nodes adapt to the query while keeping the tree itself reusable across all queries.

What would settle it

Construct a set of long documents with gold-annotated answer paragraphs, run LongRefiner twice per query—once with the predicted scope weight and once with an oracle weight chosen to maximize answer recall—and compare selection recall; if oracle-weighted selection recovers clearly more gold paragraphs, the query-scope mechanism is the bottleneck and the adaptive-refinement claim is falsified for those queries.

Watch

Extended reading notes

Core claim

At the center of the method is a document tree: each retrieved document is rewritten as a flat XML representation with section and subsection tags, a paragraph marker, and a skip token that keeps only the first and last k tokens of each block. A single 3B foundation model, trained with task-specific LoRA adapters, learns to build this tree from plain text, to classify a query as Local or Global, and to pick the globally relevant sections from the tree's outline. The paper claims that combining these three capabilities produces the best results among all compared refinement methods on seven QA datasets at a 2k-token budget, reaching or exceeding full-content accuracy on six datasets with ten times fewer tokens. The node selection rule is $\text{Score}(n_i) = \text{LS}(n_i) + R_q \cdot \text{GS}(n_i)$, where the local score propagates query-paragraph similarities upward through the tree and the global score propagates outline-based section relevance downward, with the query's predicted scope $R_q$ deciding how much weight each perspective gets.

Load-bearing premise

The load-bearing premise is that a query's predicted Local-versus-Global scope can be turned into a single numeric weight that correctly balances paragraph-level and section-level evidence; if that weight is wrong for a query, the selected nodes can omit the answer even though the document contains it.

Editorial extensions

If this is right

  • With the 2k-token budget, LongRefiner is the top-scoring method on all seven QA datasets, including multi-hop and long-form QA, while consuming roughly one-tenth the generator tokens of the full-content setting.
  • The refinement transfers across generators: switching from Llama3.1-8B-Instruct to Qwen2.5-7B-Instruct preserves the advantage over baselines, so the selected nodes are not tuned to one generator's quirks.
  • Because hierarchical structuring runs offline on the corpus, a document tree is built once and then reused for every query; the online stage reads only a few hundred input tokens and reports about 25% of the full-pipeline latency.
  • Scaling the refiner backbone from 0.5B to 7B parameters and increasing training data both improve refinement recall and downstream accuracy, meaning the approach has a predictable scaling path rather than a fixed ceiling.

Reading between the lines

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

  • The paper does not close the loop back to retrieval: the selected nodes are a natural source of positive training evidence for the retriever, and using them that way is a testable next step that follows directly from the tree-based selection.
  • Because the XML labels are derived from Wikipedia's section layout, the same pipeline on corporate or web documents with tables, navigation, and heterogeneous markup would need retraining; this is the boundary the paper itself names as its main limitation.
  • The additive weighting of local and global scores is one design choice among many; a learned gating function that takes the query text, not just a two-token softmax, might improve queries that need both scopes simultaneously, such as comparative or multi-hop questions.
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

4 major / 6 minor

Summary. The paper proposes LongRefiner, a plug-and-play document refiner for long-context retrieval-augmented generation. The method has three components: dual-level query analysis, which classifies a query as local or global and derives a continuous scope weight; hierarchical document structuring, which converts plain retrieved documents into an XML-like document tree using trainable section/subsection/paragraph tags and skip placeholders; and adaptive refinement, which scores tree nodes by combining a local reranker score with a global outline-based selection score and selects nodes until a token budget is reached. Training uses three LoRA adapters on a single Qwen2.5-3B backbone. Experiments on NQ, TriviaQA, PopQA, HotpotQA, 2WikiMultiHopQA, ASQA, and ELI5 under a 2k-token budget report that LongRefiner outperforms all compared baselines on all datasets, with a token count near 1,933 and an online latency of 10.8 seconds, and the paper claims an order-of-magnitude efficiency advantage over full-content input.

Significance. If the central results hold, LongRefiner is a practically useful contribution: it shows that structured, hierarchy-aware refinement can beat perplexity-based and retrieval-based compression methods while keeping generator input near 2k tokens, and the ablations, scaling analyses, and robustness checks across generator backbones and scoring models are valuable. The release of code and the use of external QA benchmarks for the headline accuracy numbers are strengths. However, the paper overstates the efficiency claim in the abstract, the query-adaptivity mechanism in Eq. (1) is far less adaptive than advertised, and the evaluation lacks variance reporting, so the strength of the claims exceeds what the evidence currently supports.

major comments (4)
  1. [Abstract; §4.4, Table 2] The abstract's claim of "10x fewer computational costs and latency compared to the best baseline" is not supported by Table 2. LongRefiner uses 1,933 tokens, whereas the strongest accuracy baseline, LongLLMLingua, uses 1,976 tokens, so the token reduction relative to the best baseline is about 2%, not 10x; the 10x figure only holds when comparing to the Full Content row (19,567 tokens), which is not the best baseline. The stated online latency advantage is also computed while excluding the offline hierarchical structuring stage described in Section 3.4, so the end-to-end computational cost of LongRefiner is not reported. Please rephrase the efficiency claims to specify the comparison point and report the total compute including the offline stage.
  2. [§3.1, Eq. (1); §3.3; Table 3] The adaptive query-scope weight Rq is nearly constant by construction. Since Pl and Pg in Eq. (1) are generation probabilities in [0,1], Softmax(Pl, Pg)_g lies in [1/(1+e), e/(1+e)] ≈ [0.269, 0.731] for every possible query; a strongly local query still receives at least 27% global weight and a strongly global query at most 73%. Therefore the "continuous information scope" the authors claim is compressed into less than half of the unit interval, and the degradation of the "w/o Query Analysis" ablation in Table 3 cannot be attributed to meaningful query adaptivity unless additional evidence is provided. The additive combination Score(ni) = LS(ni) + Rq * GS(ni) also adds an unnormalized reranker similarity to a diluted outline-selection score without any calibration or independent validation; this is load-bearing because every downstream selection in Table 2 depends on this sum. Please either correct the formula (e.g., using log-probabilities before softmax) or demonstrate that the resulting Rq variation actually changes node rankings in a meaningful way.
  3. [§4.4, Tables 2 and 3] The headline claim of outperforming all baselines on all seven datasets is reported as a single run with no error bars, confidence intervals, or significance tests. Greedy decoding reduces sampling variance, but not evaluation variance or sensitivity to the choice of test subset; moreover, some differences in Table 2 are very small, such as ELI5 F1 of 23.9 versus 23.7 for LongLLMLingua, and the "w/o Query Analysis" ablation in Table 3 differs by only 0.6-2.0 points depending on dataset. Since the central contribution is the claimed universal improvement, please report variance over multiple runs or a statistical significance test, and state how many examples are evaluated per dataset.
  4. [§3.3; Appendix D] The global-selection training labels are self-referential: as Appendix D states, the abstracts and outlines used to annotate the selected titles are generated by the authors' own trained structuring model rather than derived from the golden Wikipedia structures. This means the global-selection module may learn to compensate for the structuring model's systematic errors, which could inflate the module's measured contribution in the ablation and the end-to-end results. Because the main accuracy numbers are measured on external QA benchmarks, this is not a fatal circularity, but it is an unquantified bias in a load-bearing training stage. Please report an ablation in which the global-selection labels are created from gold or human-readable outlines, or provide an agreement analysis between the model-generated outlines and the ground-truth Wikipedia structure.
minor comments (6)
  1. [Footnote 1] Typo: "Correpsonding author" should be "Corresponding author."
  2. [Figure 1] Figure 1 plots "Latency" against "Performance," but the axes have no units and the identity of the plotted methods is not fully clear; please state the metric definitions and the experimental setting.
  3. [§3.2] The text says that each section's content is enclosed within <section: {title}> and </subsection: {title}>, but the closing tag should be </section: {title}> according to Table 1; this appears to be a typo.
  4. [§3.1 after Eq. (1)] The equation defines Rq, but the discussion refers to "a continuous representation rq"; please standardize the symbol to one or the other.
  5. [§3.2] The sentence "As shown in Figure 1, the XML-based Dxml reduces the token count to approximately 1/10 of the original" cites a latency-performance scatter plot that does not show token counts; please correct the reference to the relevant table or plot.
  6. [Table 4] The row "Best Baseline" is not defined; please specify which method it denotes and how it was selected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: LongRefiner's scores and predictions are not fitted to the final QA benchmarks, and the central results rest on external held-out evaluation.

full rationale

LongRefiner's derivation is self-contained against external QA benchmarks. The three learned components (query analysis, hierarchical structuring, global selection) are trained on labels that do not encode the final answer: query Local/Global labels and global selection labels are produced by Llama3.1-70B teacher prompts (Appendix D), and structuring labels come from Wikipedia webpage structure, not from downstream QA scores. The combination Score(ni) = LS(ni) + Rq * GS(ni) has no constants fitted to NQ, TriviaQA, or other test sets; LS comes from an off-the-shelf reranker and GS from the trained outline-selection module. The only self-referential choice is that global-selection training annotations use abstracts and outlines produced by the system's own structuring model rather than golden data (Appendix D). That aligns train/inference distributions and could propagate structuring errors, but it is not a reduction by construction: the final QA accuracy is measured on held-out external data, and no output of the pipeline is defined as equal to a training label or fitted value. Concerns about the narrow range of Rq under the softmax or about the unvalidated additive combination of LS and GS are correctness and calibration concerns, not circularity. Self-citations such as FlashRAG and BIDER are used only for toolkit implementation and related-work context, not to justify the central claim. No equation in the paper is identical to its input by construction, so the appropriate finding is no significant circularity.

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

The method relies on several domain assumptions about Wikipedia-derived structure, teacher-LLM labels, and the reliability of a fixed reranker. There are no explicitly fitted parameters in the scoring formula, but the hyperparameter k and an undefined multiplier g in Eq. 1 are unstated, and the token budget is an experimental choice.

free parameters (2)
  • k (first/last tokens kept per paragraph) = not reported
    Hyperparameter controlling the compression ratio and parsing errors in the <skip> representation; its value is never stated in the paper.
  • g (scaling in Eq. 1) = not reported
    Equation 1 'Rq = Softmax(Pl,Pg)g' suggests an additional multiplier on the softmax output; it is undefined and no value is given. If intentional, it is a hand-set parameter affecting how local and global scores are blended.
assumptions (4)
  • domain assumption Wikipedia structure is a valid proxy for real document structure
    Training labels for hierarchical structuring come from Wikipedia web pages; the method assumes the section/subsection/paragraph layout transfers to the retrieved documents.
  • domain assumption Teacher LLM labels are reliable
    Llama3.1-70B annotations for query type (Local/Global) and global section selection are treated as ground truth during training.
  • domain assumption Bge-reranker-v2-m3 produces reliable local relevance scores
    Local scores M(query, node) are computed with a fixed reranker, and the method assumes this ranks paragraphs correctly across all datasets.
  • domain assumption A 3B model can accurately generate XML structure for long documents
    The hierarchical structuring step is performed by Qwen2.5-3B-Instruct; the method assumes its output is accurate enough for downstream node selection.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hierarchical Document Refinement for Long-context Retrieval-augmented Generation." pith.science (2026). https://pith.science/paper/G7FHE7EO

@misc{pith2026250510413,
  author       = {Pith},
  title        = {Pith review of: Hierarchical Document Refinement for Long-context Retrieval-augmented Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G7FHE7EO}},
  note         = {Machine review of arXiv:2505.10413}
}
read the original abstract

Real-world RAG applications often encounter long-context input scenarios, where redundant information and noise results in higher inference costs and reduced performance. To address these challenges, we propose LongRefiner, an efficient plug-and-play refiner that leverages the inherent structural characteristics of long documents. LongRefiner employs dual-level query analysis, hierarchical document structuring, and adaptive refinement through multi-task learning on a single foundation model. Experiments on seven QA datasets demonstrate that LongRefiner achieves competitive performance in various scenarios while using 10x fewer computational costs and latency compared to the best baseline. Further analysis validates that LongRefiner is scalable, efficient, and effective, providing practical insights for real-world long-text RAG applications. Our code is available at https://github.com/ignorejjj/LongRefiner.

Figures

Figures reproduced from arXiv: 2505.10413 by the authors.

Figure 1
Figure 1. Comparison of different methods in terms of [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the LongRefiner Framework. sign a corresponding information level, which is represented as a binary label.1 We then treat this label as a special token and finetune the refiner to generate the corresponding special token based on the input query. During inference, LongRefiner adaptively determines the amount of information required for each query by predicting the appropri￾ate information-level token. Th… view at source ↗
Figure 4
Figure 4. The performance of LongRefiner across dif [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (1 more)
Figure 3
Figure 3. Figure 3: Analysis of scaling the base model size (left) [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Leveraging LLM-Assisted Query Understanding for Live Retrieval-Augmented Generation

    cs.CL 2025-06 conditional novelty 4.0 of 10

    Omni-RAG, a query-rewriting and decomposition pipeline on top of standard retrieval and reranking, achieved rank 2 in the SIGIR 2025 LiveRAG Challenge.

Reference graph

Works this paper leans on

56 extracted references · 8 canonical work pages · cited by 1 Pith paper

  1. [1]

    online" 'onlinestring :=

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

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Serge Abiteboul. 1999. https://doi.org/10.1145/344816.344853 On views and XML . SIGMOD Rec. , 28(4):30--38

    work page arXiv 1999

  4. [4]

    Yushi Bai, Xin Lv, Jiajie Zhang, Yuze He, Ji Qi, Lei Hou, Jie Tang, Yuxiao Dong, and Juanzi Li. 2024. https://aclanthology.org/2024.findings-emnlp.74 Longalign: A recipe for long context alignment of large language models . In Findings of the Association for Computational Linguistics: EMNLP 2024, Miami, Florida, USA, November 12-16, 2024 , pages 1376--139...

    work page 2024

  5. [5]

    Steven Bird. 2006. https://doi.org/10.3115/1225403.1225421 NLTK: the natural language toolkit . In ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006 . The Association for Computer Linguistics

    arXiv 2006

  6. [6]

    Rae, Erich Elsen, and Laurent Sifre

    Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George van den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, Diego de Las Casas, Aurelia Guy, Jacob Menick, Roman Ring, Tom Hennigan, Saffron Huang, Loren Maggiore, Chris Jones, Albin Cassirer, Andy Brock, Michela Paganini, Geoffrey Irving, Ori...

    work page 2022

  7. [7]

    Chi - Min Chan, Chunpu Xu, Ruibin Yuan, Hongyin Luo, Wei Xue, Yike Guo, and Jie Fu. 2024. https://doi.org/10.48550/ARXIV.2404.00610 RQ-RAG: learning to refine queries for retrieval augmented generation . CoRR, abs/2404.00610

    arXiv doi:10.48550/arxiv.2404.00610 2024

  8. [8]

    Yukang Chen, Shengju Qian, Haotian Tang, Xin Lai, Zhijian Liu, Song Han, and Jiaya Jia. 2024. https://openreview.net/forum?id=6PmJoRfdaK Longlora: Efficient fine-tuning of long-context large language models . In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024 . OpenReview.net

    2024

Show all 56 references
  1. [9]

    Xin Cheng, Xun Wang, Xingxing Zhang, Tao Ge, Si - Qing Chen, Furu Wei, Huishuai Zhang, and Dongyan Zhao. 2024. https://doi.org/10.48550/ARXIV.2405.13792 xrag: Extreme context compression for retrieval-augmented generation with one token . CoRR, abs/2405.13792

    2024 doi arXiv

  2. [10]

    Zhuyun Dai and Jamie Callan. 2019. https://doi.org/10.1145/3331184.3331303 Deeper text understanding for IR with contextual neural language modeling . In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2019...

    2019

  3. [11]

    Guanting Dong, Yutao Zhu, Chenghao Zhang, Zechen Wang, Zhicheng Dou, and Ji - Rong Wen. 2024. https://doi.org/10.48550/ARXIV.2406.18676 Understand what LLM needs: Dual preference alignment for retrieval-augmented generation . CoRR, abs/2406.18676

    2024 doi arXiv

  4. [12]

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  5. [13]

    Angela Fan, Yacine Jernite, Ethan Perez, David Grangier, Jason Weston, and Michael Auli. 2019. https://doi.org/10.18653/v1/P19-1346 ELI 5: Long form question answering . In ACL, pages 3558--3567, Florence, Italy. Association for Computational Linguistics

    2019 doi

  6. [14]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, Qianyu Guo, Meng Wang, and Haofen Wang. 2024. https://arxiv.org/abs/2312.10997 Retrieval-augmented generation for large language models: A survey . Preprint, arXiv:2312.10997

    2024 arXiv

  7. [15]

    Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Mingwei Chang. 2020. Retrieval augmented language model pre-training. In International conference on machine learning, pages 3929--3938. PMLR

    2020

  8. [16]

    Xanh Ho, Anh - Khoa Duong Nguyen, Saku Sugawara, and Akiko Aizawa. 2020. https://doi.org/10.18653/V1/2020.COLING-MAIN.580 Constructing A multi-hop QA dataset for comprehensive evaluation of reasoning steps . In Proceedings of the 28th International Conference on Computational ...

    2020 doi

  9. [17]

    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. https://openreview.net/forum?id=nZeVKeeFYf9 Lora: Low-rank adaptation of large language models . In The Tenth International Conference on Learning Representatio...

    2022

  10. [18]

    Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qianglong Chen, Weihua Peng, Xiaocheng Feng, Bing Qin, and Ting Liu. 2023. https://doi.org/10.48550/ARXIV.2311.05232 A survey on hallucination in large language models: Principles, taxonomy, challen...

    2023 doi arXiv

  11. [19]

    Huiqiang Jiang, Qianhui Wu, , Xufang Luo, Dongsheng Li, Chin-Yew Lin, Yuqing Yang, and Lili Qiu. 2023 a . https://arxiv.org/abs/2310.06839 Longllmlingua: Accelerating and enhancing llms in long context scenarios via prompt compression . ArXiv preprint, abs/2310.06839

    2023 arXiv

  12. [20]

    Huiqiang Jiang, Qianhui Wu, Chin-Yew Lin, Yuqing Yang, and Lili Qiu. 2023 b . https://doi.org/10.18653/v1/2023.emnlp-main.825 LLML ingua: Compressing prompts for accelerated inference of large language models . In Proceedings of the 2023 Conference on Empirical Methods in Natu...

    2023 doi

  13. [21]

    Jiajie Jin, Yutao Zhu, Xinyu Yang, Chenghao Zhang, and Zhicheng Dou. 2024 a . https://doi.org/10.48550/ARXIV.2405.13576 Flashrag: A modular toolkit for efficient retrieval-augmented generation research . CoRR, abs/2405.13576

    2024 doi arXiv

  14. [22]

    Jiajie Jin, Yutao Zhu, Yujia Zhou, and Zhicheng Dou. 2024 b . https://doi.org/10.18653/V1/2024.FINDINGS-ACL.42 BIDER: bridging knowledge inconsistency for efficient retrieval-augmented llms via key supporting evidence . In Findings of the Association for Computational Linguist...

    2024 doi

  15. [23]

    Mandar Joshi, Eunsol Choi, Daniel Weld, and Luke Zettlemoyer. 2017. T rivia QA : A large scale distantly supervised challenge dataset for reading comprehension. In ACL, pages 1601--1611, Vancouver, Canada. Association for Computational Linguistics

    2017

  16. [24]

    Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. Dense passage retrieval for open-domain question answering. In EMNLP, pages 6769--6781

    2020

  17. [25]

    Tom Kwiatkowski, Jennimaria Palomaki, Olivia Redfield, Michael Collins, Ankur Parikh, Chris Alberti, Danielle Epstein, Illia Polosukhin, Jacob Devlin, Kenton Lee, et al. 2019. Natural questions: a benchmark for question answering research. Transactions of the Association for C...

    2019

  18. [26]

    Woosuk Kwon, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph Gonzalez, Hao Zhang, and Ion Stoica. 2023. https://doi.org/10.1145/3600006.3613165 Efficient memory management for large language model serving with pagedattention . In Proceedings of the 29...

    2023

  19. [27]

    Ryan Levering and Michal Cutler. 2006. https://doi.org/10.1145/1166160.1166213 The portrait of a common HTML web page . In Proceedings of the 2006 ACM Symposium on Document Engineering, Amsterdam, The Netherlands, October 10-13, 2006 , pages 198--204. ACM

    2006

  20. [28]

    u ttler, Mike Lewis, Wen tau Yih, Tim Rockt \

    Patrick S. H. Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \" u ttler, Mike Lewis, Wen tau Yih, Tim Rockt \" a schel, Sebastian Riedel, and Douwe Kiela. 2020. https://proceedings.neurips.cc/paper/2020/hash/6b493230205f780e1b...

    2020

  21. [29]

    Xiaoxi Li, Guanting Dong, Jiajie Jin, Yuyao Zhang, Yujia Zhou, Yutao Zhu, Peitian Zhang, and Zhicheng Dou. 2025 a . https://doi.org/10.48550/ARXIV.2501.05366 Search-o1: Agentic search-enhanced large reasoning models . CoRR, abs/2501.05366

    2025 doi arXiv

  22. [30]

    Xiaoxi Li, Jiajie Jin, Guanting Dong, Hongjin Qian, Yutao Zhu, Yongkang Wu, Ji - Rong Wen, and Zhicheng Dou. 2025 b . https://doi.org/10.48550/ARXIV.2504.21776 Webthinker: Empowering large reasoning models with deep research capability . CoRR, abs/2504.21776

    2025 doi arXiv

  23. [31]

    Xiaoxi Li, Jiajie Jin, Yujia Zhou, Yuyao Zhang, Peitian Zhang, Yutao Zhu, and Zhicheng Dou. 2024 a . https://doi.org/10.48550/ARXIV.2404.14851 From matching to generation: A survey on generative information retrieval . CoRR, abs/2404.14851

    2024 doi arXiv

  24. [32]

    Yucheng Li, Bo Dong, Frank Guerin, and Chenghua Lin. 2023. https://doi.org/10.18653/V1/2023.EMNLP-MAIN.391 Compressing context to enhance inference efficiency of large language models . In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, ...

    2023 doi

  25. [33]

    Zongqian Li, Yinhong Liu, Yixuan Su, and Nigel Collier. 2024 b . https://doi.org/10.48550/ARXIV.2410.12388 Prompt compression for large language models: A survey . CoRR, abs/2410.12388

    2024 doi arXiv

  26. [34]

    Zheng Liu, Chenyuan Wu, Ninglu Shao, Shitao Xiao, Chaozhuo Li, and Defu Lian. 2024. https://doi.org/10.48550/ARXIV.2409.15699 Lighter and better: Towards flexible context adaptation for retrieval augmented generation . CoRR, abs/2409.15699

    2024 doi arXiv

  27. [35]

    Alex Mallen, Akari Asai, Victor Zhong, Rajarshi Das, Daniel Khashabi, and Hannaneh Hajishirzi. 2023. https://doi.org/10.18653/V1/2023.ACL-LONG.546 When not to trust language models: Investigating effectiveness of parametric and non-parametric memories . In Proceedings of the 6...

    2023 doi

  28. [36]

    Vicky Zhao, Lili Qiu, and Dongmei Zhang

    Zhuoshi Pan, Qianhui Wu, Huiqiang Jiang, Menglin Xia, Xufang Luo, Jue Zhang, Qingwei Lin, Victor R \" u hle, Yuqing Yang, Chin - Yew Lin, H. Vicky Zhao, Lili Qiu, and Dongmei Zhang. 2024. https://doi.org/10.18653/V1/2024.FINDINGS-ACL.57 Llmlingua-2: Data distillation for effic...

    2024 doi

  29. [37]

    Fabio Petroni, Aleksandra Piktus, Angela Fan, Patrick Lewis, Majid Yazdani, Nicola De Cao, James Thorne, Yacine Jernite, Vladimir Karpukhin, Jean Maillard, Vassilis Plachouras, Tim Rockt \"a schel, and Sebastian Riedel. 2021. https://doi.org/10.18653/v1/2021.naacl-main.200 KIL...

    2021 doi

  30. [38]

    Hongjin Qian, Zheng Liu, Kelong Mao, Yujia Zhou, and Zhicheng Dou. 2024. https://doi.org/10.18653/V1/2024.ACL-LONG.71 Grounding language model with chunking-free in-context retrieval . In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (...

    2024 doi

  31. [39]

    Nils Reimers and Iryna Gurevych. 2019. https://doi.org/10.18653/V1/D19-1410 Sentence-bert: Sentence embeddings using siamese bert-networks . In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on ...

    2019 doi

  32. [40]

    Robertson and Hugo Zaragoza

    Stephen E. Robertson and Hugo Zaragoza. 2009. https://doi.org/10.1561/1500000019 The probabilistic relevance framework: BM25 and beyond . Found. Trends Inf. Retr., 3(4):333--389

    2009 doi

  33. [41]

    Ivan Stelmakh, Yi Luan, Bhuwan Dhingra, and Ming - Wei Chang. 2022. https://doi.org/10.18653/V1/2022.EMNLP-MAIN.566 ASQA: factoid questions meet long-form answers . In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022, Abu Dhabi...

    2022 doi

  34. [42]

    Jiejun Tan, Zhicheng Dou, Yutao Zhu, Peidong Guo, Kun Fang, and Ji - Rong Wen. 2024 a . https://doi.org/10.18653/V1/2024.ACL-LONG.242 Small models, big insights: Leveraging slim proxy models to decide when and what to retrieve for llms . In Proceedings of the 62nd Annual Meeti...

    2024 doi

  35. [43]

    Patil, Ziyang Wu, Tianjun Zhang, Kurt Keutzer, Joseph Gonzalez, and Raluca A

    Sijun Tan, Xiuyu Li, Shishir G. Patil, Ziyang Wu, Tianjun Zhang, Kurt Keutzer, Joseph Gonzalez, and Raluca A. Popa. 2024 b . https://aclanthology.org/2024.emnlp-main.975 Lloco: Learning long contexts offline . In Proceedings of the 2024 Conference on Empirical Methods in Natur...

    2024

  36. [44]

    Cunxiang Wang, Xiaoze Liu, Yuanhao Yue, Xiangru Tang, Tianhang Zhang, Jiayang Cheng, Yunzhi Yao, Wenyang Gao, Xuming Hu, Zehan Qi, Yidong Wang, Linyi Yang, Jindong Wang, Xing Xie, Zheng Zhang, and Yue Zhang. 2023 a . https://doi.org/10.48550/ARXIV.2310.07521 Survey on factuali...

    2023 doi arXiv

  37. [45]

    Rizwan Parvez, and Graham Neubig

    Zhiruo Wang, Jun Araki, Zhengbao Jiang, Md. Rizwan Parvez, and Graham Neubig. 2023 b . https://doi.org/10.48550/ARXIV.2311.08377 Learning to filter context for retrieval-augmented generation . CoRR, abs/2311.08377

    2023 doi arXiv

  38. [46]

    Shitao Xiao, Zheng Liu, Peitian Zhang, Niklas Muennighoff, Defu Lian, and Jian - Yun Nie. 2024. https://doi.org/10.1145/3626772.3657878 C-pack: Packed resources for general chinese embeddings . In Proceedings of the 47th International ACM SIGIR Conference on Research and Devel...

    2024

  39. [47]

    Fangyuan Xu, Weijia Shi, and Eunsol Choi. 2023. https://doi.org/10.48550/ARXIV.2310.04408 RECOMP: improving retrieval-augmented lms with compression and selective augmentation . CoRR, abs/2310.04408

    2023 doi arXiv

  40. [48]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, Keming Lu, Keqin Bao, Kexin Yang, Le Yu, Me...

    2024 arXiv

  41. [49]

    Haoyan Yang, Zhitao Li, Yong Zhang, Jianzong Wang, Ning Cheng, Ming Li, and Jing Xiao. 2023. https://aclanthology.org/2023.emnlp-main.326 PRCA: fitting black-box large language models for retrieval question answering via pluggable reward-driven contextual adapter . In Proceedi...

    2023

  42. [50]

    Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William Cohen, Ruslan Salakhutdinov, and Christopher D. Manning. 2018. https://doi.org/10.18653/v1/D18-1259 H otpot QA : A dataset for diverse, explainable multi-hop question answering . In EMNLP, pages 2369--2380, Brussels,...

    2018 doi

  43. [51]

    Chanwoong Yoon, Taewhoo Lee, Hyeon Hwang, Minbyul Jeong, and Jaewoo Kang. 2024. https://aclanthology.org/2024.emnlp-main.1194 Compact: Compressing retrieved documents actively for question answering . In Proceedings of the 2024 Conference on Empirical Methods in Natural Langua...

    2024

  44. [52]

    Jihao Zhao, Zhiyuan Ji, Pengnian Qi, Simin Niu, Bo Tang, Feiyu Xiong, and Zhiyu Li. 2024. https://doi.org/10.48550/ARXIV.2410.12788 Meta-chunking: Learning efficient text segmentation via logical perception . CoRR, abs/2410.12788

    2024 doi arXiv

  45. [53]

    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...

    2023 doi arXiv

  46. [54]

    Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, Zheyan Luo, and Yongqiang Ma. 2024. https://doi.org/10.48550/ARXIV.2403.13372 Llamafactory: Unified efficient fine-tuning of 100+ language models . CoRR, abs/2403.13372

    2024 doi arXiv

  47. [55]

    Yujia Zhou, Yan Liu, Xiaoxi Li, Jiajie Jin, Hongjin Qian, Zheng Liu, Chaozhuo Li, Zhicheng Dou, Tsung - Yi Ho, and Philip S. Yu. 2024. https://doi.org/10.48550/ARXIV.2409.10102 Trustworthiness in retrieval-augmented generation systems: A survey . CoRR, abs/2409.10102

    2024 doi arXiv

  48. [56]

    Wenhu Chen Ziyan Jiang, Xueguang Ma. 2024. https://arxiv.org/abs/2406.15319 Longrag: Enhancing retrieval-augmented generation with long-context llms . arXiv preprint arXiv:2406.15319

    2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.