Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

MacRAG: Compress, Slice, and Scale-up for Multi-Scale Adaptive Context RAG

T0 review · 4 major / 5 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 MacRAG claims that starting retrieval at fine summary slices and expanding to parent chunks and neighboring text improves multi-hop RAG accuracy across three benchmarks and three LLMs.

desk verdict A solid, well-engineered RAG pipeline with unusually broad evaluation, but the paper's own tables contradict the 'consistently surpasses' claim, and the missing significance testing makes the central result provisional. read the letter →

arxiv 2505.06569 v2 pith:L7XYLX32 submitted 2025-05-10 cs.CL cs.AIcs.IRcs.LG

Woosang Lim , Zekun Li , Gyuwan Kim , Sungyoung Ji , HyeonJung Kim , Kyuri Choi , Jin Hyuk Lim , Kyungpyo Park
show 1 more author
William Yang Wang
This is my paper · ORCID
classification cs.CLcs.AIcs.IRcs.LG
keywords retrieval-augmentedgenerationmulti-hopquestionansweringhierarchicalretrievallong-contextlanguagemodelscontextcompressionmulti-scaleLongBenchadaptiveconstruction
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

MacRAG is a retrieval pipeline designed to fix a specific failure: standard RAG either retrieves chunks too coarse to be precise or too fine to carry the connecting facts needed for multi-hop questions. The paper claims that a hierarchical index, documents split into chunks, each chunk compressed into a summary, each summary sliced into small overlapping units, lets retrieval start at the finest level and then expand upward to parent chunks, neighboring chunks, and entire documents. This compress, slice, and scale-up order, the paper argues, gives both precision and coverage while keeping the final context bounded. Evaluated on LongBench versions of HotpotQA, 2WikiMultihopQA, and Musique, MacRAG consistently beats RAPTOR and LongRAG in F1 across Llama-3.1-8B, Gemini-1.5-pro, and GPT-4o in both single-step and multi-step generation. If true, this would make multi-hop QA more accurate without simply feeding the model more tokens.

What carries the argument

The load-bearing object is a four-level hierarchy, document to chunk to summary to slice, built once offline, combined with a bottom-up retrieval procedure that walks back up that hierarchy at query time. Each document is split into overlapping chunks of roughly 200 to 500 tokens; each chunk is abstractively summarized; and each summary is split into overlapping slices of roughly 50 to 200 tokens that are embedded and stored with metadata linking each slice to its parent chunk and document. Retrieval then runs in five steps: slice-level top-$k_1$ search, unique parent-chunk mapping, cross-encoder chunk reranking, scaled top-($k_2 \times \alpha$) chunk selection followed by document ranking to pick $k_2$ distinct documents, and finally merging each chosen chunk with its $h$-hop neighbors to form bounded final contexts. The mechanism's work is to let the cheap, precise slice search decide where to look and then let the cheap index-based lookups decide how much context to add, so the model sees a query-specific, length-bounded assembly rather than whole documents or a static chunk list.

What would settle it

Run MacRAG on HotpotQA, 2WikiMultihopQA, and Musique with the summarizer replaced by an identity function, retrieving raw slices instead of summarized slices while keeping chunking, slicing, reranking, and neighbor merging fixed; if F1 stays roughly the same, compression is not load-bearing. Alternatively, check whether all gold evidence passages referenced by the dataset questions survive the summarization step; if a large fraction of bridging entities or relations disappear from the summaries, the claimed gains cannot be attributed to preserving facts through compression.

Watch

Extended reading notes

Core claim

Starting from the finest retrieval granularity and expanding only as needed is enough to lift multi-hop reasoning accuracy above existing hierarchical RAG baselines. Concretely, MacRAG first retrieves the top $k_1$ summary slices by query similarity, maps them to their parent chunks, reranks those chunks with a cross-encoder, widens the candidate set by a factor $\alpha$ to catch borderline chunks, ranks the source documents, and finally merges each selected chunk with its $h$-hop neighbors to form the final top-$k_2$ merged contexts. On the LongBench versions of HotpotQA, 2WikiMultihopQA, and Musique, this yields average F1 gains over LongRAG of about +2.05 with Llama-3.1-8B, +6.10 with Gemini-1.5-pro, and +5.29 with GPT-4o when paired with LongRAG's extract-and-filter generation, and a +2.37 average F1 gain over RAPTOR with Llama-3.1-8B; the gains are largest on Musique, the dataset the paper identifies as the hardest retrieval setting. The paper also reports a 7.2% precision gain from indexing summaries rather than raw slices, sub-second retrieval and reranking at 0.23 seconds on average, and ablation results in which removing the propagation-and-merging step costs up to about 5% F1 and removing the scaling-up step costs up to about 6%.

Load-bearing premise

The pipeline assumes that compressing each chunk into a summary and then slicing the summary preserves the specific facts, the entities and relations, needed for multi-hop reasoning, because retrieval only ever sees summarized slices; if summarization drops a bridging fact, no later expansion can recover it.

Editorial extensions

If this is right

  • Multi-hop QA systems can raise F1 by starting retrieval at summary slices and expanding to parent chunks and neighbors, rather than by retrieving whole documents or fixed chunks.
  • The gains persist across an 8B open model and two large proprietary models, so the retrieval architecture's benefit is largely complementary to model scale.
  • Because the final context is bounded to $k_2$ merged chunks, MacRAG keeps retrieval and reranking near 0.23 seconds per query, and its intermediate generation variant can beat LongRAG's full-document extraction while using less context.
  • The method drops by roughly 5% F1 when neighbor propagation is removed and roughly 6% when scaling-up is removed, so both expansion steps are load-bearing for multi-hop coverage.
  • MacRAG's retrieved contexts can feed seven different single- and multi-step generation schemes, so it behaves as a drop-in retrieval module rather than a single end-to-end recipe.

Reading between the lines

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

  • A testable extension: measure summary fidelity directly by checking whether gold evidence entities and relations from each question survive the chunk-to-summary compression; if they often do not, the reported gains would have to come from the expansion steps rather than from compression preserving facts.
  • The $h$-hop neighbor merge suggests a cheap alternative to explicit graph construction: linear adjacency in document order may capture many bridging relations that graph-based indexes are built to find, at lower indexing cost.
  • Because the final context size is controlled by $k_2$ and $\alpha$, MacRAG looks well suited as a context-assembly module inside iterative or agentic RAG loops, where each step's token budget must stay stable; the paper proposes this direction but does not evaluate it.
  • The 7.2% precision gain from summaries over raw slices is preliminary; a controlled sweep over summarization models and compression ratios would reveal whether abstractive compression is the load-bearing ingredient or whether slicing alone drives most of the gain.
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 / 5 minor

Summary. MacRAG is a hierarchical RAG framework that builds an offline index by compressing each chunk into a summary, slicing the summary into fine-grained units, and then at query time retrieving slices, mapping them to parent chunks, reranking with a cross-encoder, scaling up candidate chunks, selecting top documents, and merging h-hop neighbor chunks into a bounded final context. The paper evaluates MacRAG on LongBench versions of HotpotQA, 2WikiMultihopQA, and Musique using Llama-3.1-8B, Gemini-1.5-pro, and GPT-4o under seven generation modes, comparing against RAPTOR and LongRAG. The headline claim is that MacRAG consistently surpasses baseline RAG pipelines in single- and multi-step generation.

Significance. If substantiated, MacRAG would be a useful modular contribution: it is simple, gives positive average F1 gains over LongRAG in most settings, keeps the final context bounded, and ships with released code and matched reranker/hyperparameter controls that mitigate the usual comparison confounds. The gains on Musique and 2WikiMultihopQA with stronger LLMs are particularly encouraging. However, the paper's central claim of 'consistent' superiority is stronger than the evidence: there are multiple negative cells, no paired significance testing, and an ablation narrative that is contradicted by parts of the ablation table. With corrected claims and uncertainty quantification, this can become a solid empirical study, but in its current form the main conclusion is overstated.

major comments (4)
  1. [Section 4.2, Table 3] The ablation narrative is not supported by the reported numbers. The text states that 'each component contributes to cumulative performance gains' and that removing Propagation&Merging causes drops of up to about 5 F1, but Table 3 shows improvements from removing this component in several cells, e.g., Musique Full_E&F (52.26 vs 50.09, +2.17), 2Wiki R&B_E&F (72.43 vs 71.40, +1.03), and HotpotQA Fil (68.06 vs 65.44, +2.62). This directly contradicts the claim that removing propagation and merging hurts performance and needs to be reconciled, either by revising the text or by explaining why these cells behave differently.
  2. [Abstract, Section 4.2, Tables 6 and 7] The claim that MacRAG 'consistently surpasses' LongRAG and RAPTOR is contradicted by the paper's own results. For example, Table 6 shows MacRAG behind LongRAG on HotpotQA R&B (63.02 vs 63.59) and on 2Wiki R&B (58.38 vs 60.13) with Gemini-1.5-pro, and Table 1 shows MacRAG at 44.87 F1 on 2Wiki with Llama-3.1-8B, below the plain RAG-with-reranking baseline's 46.33. Table 7 adds further negative cells, such as HotpotQA R&B with bge-m3 at k2=7 (-0.08) and Musique R&L with bge-m3 at k2=7 (-1.40). No confidence intervals, error bars, or paired significance tests are reported for any of these comparisons, so the evidence does not support the strong wording in the abstract and Section 4.2. The authors should either provide significance testing, or soften the central claim to positive average gains with explicitly quantified variance.
  3. [Section 3.1.2, Section 3.2.1, Section 4.2] The core retrieval assumption, that compressing each chunk into a summary and then retrieving only from slices of summaries preserves the facts needed for multi-hop reasoning, is never verified. If the summarizer drops a bridging entity or relation, the later parent-chunk and document expansions cannot recover it, because slice-level retrieval only sees summarized content. The preliminary 7.2% precision gain from summaries over raw slices reported in Section 4.2 does not measure recall or evidence fidelity. A direct evaluation of whether gold supporting evidence survives compression (e.g., passage-level or entity-level recall on the retrieved slices) is needed to establish that the compression step is not a hidden source of information loss.
  4. [Section 4.2, Table 5] The sentence 'regardless of the choice of k2 or reranker, MacRAG maintains substantial advantages in all test settings in Table 5' is not accurate. Table 5 contains negative cells, including HotpotQA R&B with bge-m3 at k2=7 (-0.08) and Musique R&L with bge-m3 at k2=7 (-1.40). This internal inconsistency between the prose and the table must be fixed, and the robustness claim should be limited to the settings where it actually holds.
minor comments (5)
  1. [Section 3.3, Tables 6 and 7] The generation-mode naming is inconsistent: Tables 6 and 7 use 'R&B_Ext_Fil' while the text and Table 3 consistently use 'R&B_E&F' for what appears to be the same mode; unify the terminology.
  2. [Table 1] The gain notation in Table 1 is ambiguous because different rows compare against different baselines within the same table; add an explicit legend or baseline column so the reader can tell whether a parenthetical is relative to RAPTOR, LongRAG, or the plain RAG-with-reranking row.
  3. [Table 7, caption] The caption describes the system as 'MacRAG+LongRAG' while all other tables and the text call it 'MacRAG'; clarify whether this is the same method or a distinct pipeline.
  4. [Section 4.2, Appendix A.1] The text refers to 'Table 4 in Appendix A.1' but Table 4 appears in the main body, and the appendix reference is confusing; use consistent numbering and placement.
  5. [Figure 3] On the x-axis of Figure 3, the value labeled 'max' is not defined; state what alpha value or selection criterion this corresponds to.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: MacRAG's central claims are empirical comparisons against external baselines, with no fitted parameter disguised as a prediction and no load-bearing self-citation chain.

full rationale

The paper's central claim is an empirical one: MacRAG outperforms LongRAG and RAPTOR on multi-hop QA benchmarks across several LLMs, rerankers, and generation schemes. Nothing in the architecture is defined in terms of the evaluation outcome. The chunk sizes, slice sizes, scale-up factor alpha, hop count h, and retrieval counts k1/k2 are fixed configuration choices; the text states that LongRAG's reported optimal hyperparameters (k1=100, k2=7) are reused for both systems, and MacRAG-specific parameters are ablated rather than tuned against the test results. The compression step is a design assumption about summarization preserving evidence, but it is not derived from the target result and is not an input-output tautology. The paper does cite prior work (e.g., SIRERAG's comparison table) to justify focusing on RAPTOR and LongRAG, but that selection rationale is not load-bearing for the main comparative evaluation, which is run directly against those baselines with matched rerankers and hyperparameters. No uniqueness theorem, ansatz smuggled via self-citation, or renaming of a known result is present. The abstract's word 'consistently' is arguably overstrong given several negative cells in Tables 3, 5, 6, and 7, and the absence of uncertainty quantification is a legitimate soundness concern, but that concern concerns statistical support and framing, not circularity. Since no specific reduction of a claimed result to its inputs can be exhibited, the appropriate circularity score is 0.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the fidelity of the offline summarization and slicing stage, and on the assumption that vector-similarity ranking on slices transfers to chunks. These are plausible domain assumptions, but they are not verified independently in the paper. The free parameters are all standard RAG hyperparameters, mostly inherited from the baseline; none are fitted to the test data.

free parameters (6)
  • k1 (slice retrieval count) = 100
    Inherited from LongRAG's reported optimal hyperparameter; controls how many slices are initially retrieved.
  • k2 (final chunk count) = 7 (also 12 in robustness tests)
    Inherited from LongRAG's reported best; number of merged chunks passed to the generator.
  • alpha (scale-up factor) = 1 or 4 in main experiments; 2,3,4 in ablation
    Chosen to include borderline candidate chunks; the paper argues it is robust, but the value is a hand-set hyperparameter, not derived.
  • h (neighbor hop count) = 0 or 1
    Controls how many neighboring chunks are merged; fixed by hand.
  • chunk size and overlap = about 400 tokens per chunk; 10-200 token overlap
    Selected from the range used in prior work; impacts summary fidelity and retrieval granularity.
  • slice size and overlap = 450-600 characters with 300-450 character overlaps
    Hand-tuned to balance precision and index size, reported in Appendix A.5.
assumptions (4)
  • domain assumption Abstractive summarization of a chunk retains the core factual information needed for later retrieval.
    Section 3.1.2 introduces 'Compress' and relies on it to reduce redundancy while preserving facts; no fidelity measurement is provided.
  • domain assumption Dense vector similarity between query and summary slices identifies the correct parent chunks.
    Section 3.2.1 selects top-k1 slices by relevance score; this assumes the embedding model ranks gold evidence highly.
  • domain assumption A cross-encoder reranker on the parent chunk improves over slice-level scores.
    Section 3.2.3 assumes chunk-level reranking mitigates information fragmentation.
  • domain assumption The neighbor-merge operation does not insert harmful irrelevant content.
    Section 3.2.5 merges h-hop neighbors to add context, assuming those neighbors are relevant to the query evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MacRAG: Compress, Slice, and Scale-up for Multi-Scale Adaptive Context RAG." pith.science (2026). https://pith.science/paper/L7XYLX32

@misc{pith2026250506569,
  author       = {Pith},
  title        = {Pith review of: MacRAG: Compress, Slice, and Scale-up for Multi-Scale Adaptive Context RAG},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L7XYLX32}},
  note         = {Machine review of arXiv:2505.06569}
}
read the original abstract

Long-context large language models (LC LLMs) combined with retrieval-augmented generation (RAG) hold strong potential for complex multi-hop and large-document tasks. However, existing RAG systems often suffer from imprecise retrieval, incomplete context coverage under constrained windows, and fragmented information from suboptimal context construction. We introduce Multi-scale Adaptive Context RAG (MacRAG), a hierarchical RAG framework that compresses and partitions documents into coarse-to-fine granularities, then adaptively merges relevant contexts through real-time chunk- and document-level expansions. By initiating with finest-level retrieval and progressively incorporating broader, higher-level context, MacRAG constructs effective query-specific long contexts, optimizing both precision and coverage. Evaluations on challenging LongBench expansions of HotpotQA, 2WikiMultihopQA, and Musique confirm MacRAG consistently surpasses baseline RAG pipelines in single- and multi-step generation using Llama-3.1-8B, Gemini-1.5-pro, and GPT-4o. Our results establish MacRAG as an efficient, scalable solution for real-world long-context, multi-hop reasoning. Our code is available at https://github.com/Leezekun/MacRAG.

Figures

Figures reproduced from arXiv: 2505.06569 by the authors.

Figure 1
Figure 1. An overview of the MacRAG framework, consisting of two main phases: (1) top-down hierarchical [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Performance comparison of GPT-4o using LongRAG and MacRAG across the seven settings in Table [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Performance trends across datasets for scale factors ( [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Cumulative LLM’s input context lengths for [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Performances of LongRAG and MacRAG regarding the fours metrics (Exact Match, F1-score, Precision, [PITH_FULL_IMAGE:figures/full_fig_p014_5.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. CoinRAG: Contextualized Information Nugget KV Cache Reuse for Long-Context RAG

    cs.CL 2026-08 conditional novelty 6.0 of 10

    CoinRAG reuses offline-computed, nugget-granular KV caches to make retrieval-augmented generation faster and more accurate under a 100 ms prefill budget.

Reference graph

Works this paper leans on

25 extracted references · 5 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]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2024. https://openreview.net/forum?id=hSyW5go0v8 Self- RAG : Learning to retrieve, generate, and critique through self-reflection . In The Twelfth International Conference on Learning Representations

    2024

  4. [4]

    Yushi Bai, Xin Lv, Jiajie Zhang, Hongchang Lyu, Jiankai Tang, Zhidian Huang, Zhengxiao Du, Xiao Liu, Aohan Zeng, Lei Hou, et al. 2023. Longbench: A bilingual, multitask benchmark for long context understanding. arXiv preprint arXiv:2308.14508

    arXiv 2023

  5. [5]

    Yiqun Chen, Lingyong Yan, Weiwei Sun, Xinyu Ma, Yi Zhang, Shuaiqiang Wang, Dawei Yin, Yiming Yang, and Jiaxin Mao. 2025. https://arxiv.org/abs/2501.15228 Improving retrieval-augmented generation through multi-agent reinforcement learning . Preprint, arXiv:2501.15228

    arXiv 2025

  6. [7]

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

    arXiv 2024

  7. [8]

    Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. 2024. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130

    arXiv 2024

  8. [9]

    Bernal Jim \'e nez Guti \'e rrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. 2024. Hipporag: Neurobiologically inspired long-term memory for large language models. In The Thirty-eighth Annual Conference on Neural Information Processing Systems

    work page 2024

Show all 25 references
  1. [10]

    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

  2. [11]

    Soyeong Jeong, Jinheon Baek, Sukmin Cho, Sung Ju Hwang, and Jong C Park. 2024. Adaptive-rag: Learning to adapt retrieval-augmented large language models through question complexity. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Comp...

    2024

  3. [12]

    Huiqiang Jiang, Qianhui Wu, Chin-Yew Lin, Yuqing Yang, and Lili Qiu. 2023. Llmlingua: Compressing prompts for accelerated inference of large language models. arXiv preprint arXiv:2310.05736

    2023 arXiv

  4. [13]

    Quinn Leng, Jacob Portes, Sam Havens, Matei Zaharia, and Michael Carbin. 2024. Long context rag performance of large language models. arXiv preprint arXiv:2411.03538

    2024 arXiv

  5. [14]

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

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \"u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \"a schel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Proc...

    2020

  6. [15]

    Nelson F Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024. Lost in the middle: How language models use long contexts. Transactions of the Association for Computational Linguistics, 12:157--173

    2024

  7. [16]

    OpenAI. 2024. https://openai.com/index/hello-gpt-4o Hello gpt-4o . OpenAI Blogs

    2024

  8. [17]

    Parth Sarthi, Salman Abdullah, Aditi Tuli, Shubh Khanna, Anna Goldie, and Christopher D Manning. 2024. Raptor: Recursive abstractive processing for tree-organized retrieval. In The Twelfth International Conference on Learning Representations

    2024

  9. [18]

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, et al. 2024. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context. arXiv preprint arXiv:2403.05530

    2024 arXiv

  10. [19]

    Liang Wang, Haonan Chen, Nan Yang, Xiaolong Huang, Zhicheng Dou, and Furu Wei. 2025. https://arxiv.org/abs/2501.14342 Chain-of-retrieval augmented generation . Preprint, arXiv:2501.14342

    2025

  11. [20]

    Fangyuan Xu, Weijia Shi, and Eunsol Choi. 2024 a . Recomp: Improving retrieval-augmented lms with context compression and selective augmentation. In The Twelfth International Conference on Learning Representations

    2024

  12. [21]

    Peng Xu, Wei Ping, Xianchao Wu, Lawrence McAfee, Chen Zhu, Zihan Liu, Sandeep Subramanian, Evelina Bakhturina, Mohammad Shoeybi, and Bryan Catanzaro. 2024 b . Retrieval meets long context large language models. In The Twelfth International Conference on Learning Representations

    2024

  13. [22]

    Shi-Qi Yan, Jia-Chen Gu, Yun Zhu, and Zhen-Hua Ling. 2024. Corrective retrieval augmented generation. arXiv preprint arXiv:2401.15884

    2024 arXiv

  14. [23]

    Tan Yu, Anbang Xu, and Rama Akkiraju. 2024. In defense of rag in the era of long-context language models. arXiv preprint arXiv:2409.01666

    2024 arXiv

  15. [24]

    Zhenrui Yue, Honglei Zhuang, Aijun Bai, Kai Hui, Rolf Jagerman, Hansi Zeng, Zhen Qin, Dong Wang, Xuanhui Wang, and Michael Bendersky. 2024. Inference scaling for long-context retrieval augmented generation. arXiv preprint arXiv:2410.04343

    2024 arXiv

  16. [25]

    Nan Zhang, Prafulla Kumar Choubey, Alexander Fabbri, Gabriel Bernadett-Shapiro, Rui Zhang, Prasenjit Mitra, Caiming Xiong, and Chien-Sheng Wu. 2025. https://openreview.net/forum?id=yp95goUAT1 Sire RAG : Indexing similar and related information for multihop reasoning . In The T...

    2025

  17. [26]

    Qingfei Zhao, Ruobing Wang, Yukuo Cen, Daren Zha, Shicheng Tan, Yuxiao Dong, and Jie Tang. 2024. Longrag: A dual-perspective retrieval-augmented generation paradigm for long-context question answering. In Proceedings of the 2024 Conference on Empirical Methods in Natural Langu...

    2024

Pith tools

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