Pith. sign in

REVIEW 4 major objections 6 minor 46 references

Optimizing Question Semantic Space for Dynamic Retrieval-Augmented Multi-hop Question Answering

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Q-DREAM claims that splitting multi-hop questions into subquestions, rewriting dependent steps, and retrieving each subquestion in a cluster-specific embedding space produces state-of-the-art accuracy on multi-hop QA benchmarks and…

desk verdict A plausible modular RAG pipeline whose headline SOTA claim is undercut by a trained-versus-zero-shot comparison and test-set tuning of k and N. read the letter →

arxiv 2506.00491 v1 pith:LANHTRVQ submitted 2025-05-31 cs.IR

Linhao Ye , Lang Yu , Zhikai Lei , Qin Chen , Jie Zhou , Liang He This is my paper
classification cs.IR
keywords retrieval-augmentedgenerationmulti-hopquestionansweringdecompositionsemanticmatchingcluster-specificretrievalLoRAout-of-domaingeneralizationefficiency
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

Multi-hop questions like "Who is the child of the director of Mukhyamantri (1996 Film)?" require several retrieval steps, and conventional RAG often pulls out passages that are semantically close but not actually useful. Q-DREAM addresses this by decomposing the question, rewriting any subquestion that depends on an earlier answer into a self-contained query, and retrieving each subquestion through a separately tuned embedding space chosen by its semantic cluster. The paper claims this pipeline outperforms existing RAG baselines with both ChatGPT and Llama2-7B backbones, generalizes from 2WikiMQA to HotpotQA and IIRC without retraining, and runs about six times faster per question than interleaved retrieval-and-reasoning. If true, it would mean that organizing retrieval by question type can buy both accuracy and speed for multi-step open-domain questions.

What carries the argument

The load-bearing mechanism is the Dynamic Passage Retrieval Module (DPRM), which replaces one shared retriever with $k$ cluster-specific retrievers. Subquestions are embedded and partitioned with $k$-means; for each cluster $C_i$, a LoRA adapter is fine-tuned to maximize the cosine similarity $\cos(E_q, E_p)$ between a subquestion $q$ and its labeled helpful passage $p$ (the paper's Eq. 3). At inference, a subquestion is assigned to the nearest cluster, and only that cluster's LoRA block encodes the question and the candidate passages before the highest-scoring passage is selected. The Subquestion Dependency Optimizer (SDOM) supplies DPRM with complete queries by rewriting markers such as '#1#' using the retrieved passage for the earlier subquestion, and the Question Decomposition Module (QDM) provides the initial subquestions; together they convert a multi-hop problem into a sequence of typed single-hop retrievals.

What would settle it

The central claim would be falsified if, on the same QDM and SDOM subquestions, a single shared embedding encoder matched or beat DPRM's retrieval recall on out-of-domain data, since the paper attributes the gain specifically to the cluster-specific spaces; a direct test would compare retrieval recall on HotpotQA and IIRC with and without the per-cluster LoRA blocks.

Watch

Extended reading notes

Core claim

The paper's central claim is that the bottleneck in multi-hop retrieval-augmented QA is semantic mismatching: questions and their genuinely helpful passages sit in different neighborhoods of a shared embedding space, so a single retriever favors plausible but useless text. Q-DREAM's response is to make retrieval dynamic and typed. The Question Decomposition Module (QDM) splits the original question into subquestions; the Subquestion Dependency Optimizer (SDOM) turns dependent subquestions such as 'Who is the child of the director from #1#?' into complete queries using the earlier subquestion's retrieved passage; and the Dynamic Passage Retrieval Module (DPRM) assigns each subquestion to one of $k$ semantic clusters, each with its own LoRA-tuned embedding space that pulls helpful passages close to questions sharing the same relational pattern. Trained on 2WikiMQA and evaluated with ChatGPT and Llama2-7B, Q-DREAM reports exact-match/F1 of 48.6/62.1 on 2WikiMQA, 48.4/60.9 on HotpotQA, and 28.2/31.9 on IIRC, with retrieval precision 81.8 and recall 85.7 on 2WikiMQA, gains the paper attributes to the cluster-specific embedding spaces rather than to the backbone model.

Load-bearing premise

The method assumes that every training subquestion has a known 'helpful passage' to pull toward, and that clusters learned on 2WikiMQA will assign unseen subquestions to LoRA blocks that align them with genuinely relevant passages.

Editorial extensions

If this is right

  • A multi-hop question can be answered in one pass of decomposition, dependency rewriting, and per-subquestion retrieval, so the generator never needs to interleave its own reasoning with retrieval calls.
  • Retrieval precision and recall rise sharply when each semantic cluster of subquestions has its own LoRA-tuned embedding space: on 2WikiMQA, Q-DREAM reports 81.8 precision and 85.7 recall versus 53.8 and 68.9 for IRCoT.
  • Because the modules are trained once on 2WikiMQA and then applied without retraining, the same pipeline transfers directly to HotpotQA and IIRC.
  • The pipeline works with a small generator: Q-DREAM with Llama2-7B beats RAG baselines built on ChatGPT, suggesting the gain comes from retrieval organization rather than model scale.
  • Inference is about six times faster than IRCoT (4 versus 25 seconds per sample on average), because the framework avoids multiple round trips between retriever and generator.

Reading between the lines

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

  • Going beyond the paper: the same cluster-indexed retrieval scheme should transfer to any domain with stable relational patterns and varying entities, such as 'mayor of city' or 'CEO of company', because DPRM learns the pattern rather than the specific entities.
  • Going beyond the paper: the gold-passage requirement could be relaxed by using an existing retriever to generate pseudo-helpful passages and keeping only those that lead to a correct final answer, which would let DPRM be trained on new domains without annotated passages.
  • Going beyond the paper: the reported speedup is measured against interleaved retrieval-and-reasoning; comparing against a decomposition-only baseline with a shared encoder would isolate how much of the efficiency gain comes from avoiding round trips versus from the per-cluster retrieval setup.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes Q-DREAM, a three-module pipeline for retrieval-augmented multi-hop question answering. The Question Decomposition Module (QDM) splits a multi-hop question into subquestions; the Subquestion Dependency Optimizer Module (SDOM) rewrites dependent subquestions using previously retrieved passages; and the Dynamic Passage Retrieval Module (DPRM) clusters subquestions and applies cluster-specific LoRA adapters to an embedding model so that question and helpful-passage embeddings are aligned in separate semantic spaces. The modules are trained on 13,363 samples from 2WikiMQA, with QDM/SDOM labels generated by ChatGPT and DPRM trained with helpful-passage supervision. Experiments on 2WikiMQA, HotpotQA, and IIRC report EM/F1 improvements over InstructRAG, ChatQA2, ChatGPT, SURE, and IRCoT, plus a retrieval efficiency comparison against IRCoT and retrieval accuracy comparisons against BM25, Contriever, and baseline retrievers.

Significance. If the results are reproducible and the comparison is fair, Q-DREAM would be a useful contribution: it addresses a real limitation of one-step RAG for multi-hop QA, the architecture is model-agnostic, the DPRM idea of cluster-specific retrieval spaces is interesting, and the paper includes ablation studies showing that each module contributes positively (Table 2). The efficiency claim, if validated, would also be practically important. However, the empirical claims as presented are not yet established: the main comparison confounds architecture with supervised training on the in-domain dataset, and the hyperparameters k and N appear to be tuned on the test sets. These issues are load-bearing for the central state-of-the-art claim, so the paper needs a substantial revision of the experimental protocol before the results can be accepted.

major comments (4)
  1. [§4.1–§4.2, Table 1] The central SOTA claim is not supported by a fair comparison. Q-DREAM is trained on 13,363 samples from 2WikiMQA (§4.3), covering QDM, SDOM, and DPRM, whereas the baselines in Table 1 are not described as receiving any 2WikiMQA training: IRCoT and SURE are zero-shot prompting methods, and InstructRAG/ChatQA2 are general-purpose models. The reported gains on 2WikiMQA could therefore reflect supervised training on the test distribution, and the HotpotQA/IIRC gains could reflect transfer from 2WikiMQA supervision rather than the proposed architecture. Please report results for baselines trained on the same 2WikiMQA data, or provide an untrained variant of Q-DREAM's modules, and state clearly which systems use 2WikiMQA supervision.
  2. [§5.3, Figures 3 and 4] The hyperparameters N (number of retrieved passages per subquestion) and k (number of semantic clusters) appear to be selected from test-set accuracy curves in Figures 3 and 4. If the numbers in Table 1 use the per-dataset optimal N and k, then part of the reported EM/F1 is fitted to the evaluation sets, inflating the in-domain and out-of-domain results. Please select these hyperparameters on a held-out validation split disjoint from the reported test sets, or report the selection procedure and show results for fixed N/k across all datasets.
  3. [§5.3, Table 4 and Eq. (3)] The retrieval comparison is difficult to interpret because DPRM is trained with the 'helpful passage' supervision in Eq. (3), while BM25, Contriever, and the retrieval components of the RAG baselines are not trained on the same gold-passage labels. It is expected that a retriever trained on helpful-passage pairs would achieve higher precision and recall on 2WikiMQA. Please clarify the retrieval evaluation protocol, including whether the gold helpful passages are used as training labels, evaluation targets, or both, and report a DPRM variant trained only on the question-passage signals available to the IR baselines. This is necessary to support the claim of improved retrieval accuracy.
  4. [§5.3, Table 3] The efficiency claim rests on a single comparison against IRCoT, and the three IRCoT runs vary between 8,887 s and 16,138 s (Table 3), which suggests uncontrolled conditions such as API variance, caching effects, or concurrent load. As reported, the '6× faster' conclusion is not reliable. Please describe the measurement setup in detail, report the number of samples and per-run variance, and compare against at least one additional retrieval-augmented baseline under identical conditions.
minor comments (6)
  1. [Abstract and §1] The abstract contains a typo: 'Whereas,conventional' should be 'Whereas, conventional'; in §1, 'root temperature' should be 'room temperature'.
  2. [§3.2, Eq. (3)] In Eq. (3), the notation 'E(q,p)∈Ci log ...' is unclear; it should be a sum over the pairs in cluster Ci, e.g., '∑_{(q,p)∈Ci} log ...'.
  3. [§3.2] The text refers to 'SDQM' in 'subquestions processed through QDM and SDQM'; this should be 'SDOM'.
  4. [Algorithm 1, line 11] The notation 'E_{q_i*} = h^{(M)}_{Last} ← M^{c_{q_i*}}_{DPRM}(q_i*)' is confusing; it mixes an assignment with the extraction of the last hidden state. Please rewrite this line to separate the forward pass from the hidden-state extraction.
  5. [Table 1] The table reports single EM/F1 values without error bars or significance tests; given the small margins in some comparisons, please provide variance or significance information.
  6. [Limitations] The Limitations section does not mention the evaluation-fairness issues (training on 2WikiMQA, test-set hyperparameter selection) or the reliance on ChatGPT-generated decomposition labels; these should be acknowledged.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Q-DREAM's reported gains are empirical measurements, not derivations from fitted parameters or self-citations.

full rationale

Q-DREAM is an empirical retrieval pipeline; none of its claimed results is obtained by defining an input in terms of the output. DPRM's training objective (Eq. 3) maximizes the cosine similarity between subquestions and labeled helpful passages within k-means clusters; this is supervised alignment, and the retrieved passages used in evaluation are selected by cosine similarity (Eq. 6, Algorithm 1), not by reusing the training labels. QDM and SDOM are trained with standard likelihood objectives (Eqs. 1-2). In-domain evaluation trains on 2WikiMQA and tests on the same benchmark's test split; out-of-domain evaluation transfers to HotpotQA and IIRC, so those numbers are genuine transfer results rather than identities. The one self-citation (Ye et al., 2024) is in the related-work survey and is not load-bearing. The selection of N and k from the test-set curves in Figures 3 and 4 and the lack of a trained-baseline comparison are fairness/validity concerns about the empirical comparison, not circularity: the reported EM/F1 are not algebraically forced by those hyperparameters. No uniqueness theorem, ansatz-by-citation, or renaming of a known result appears. The Limitations section raises generalization concerns, not circular ones. Under the rule that only definitional or self-citation-based reductions count, the paper shows no material circularity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical or ontological entities. The cluster-specific LoRA blocks are model parameters; the semantic spaces are mathematical constructions. The main unpaid assumptions are the transferability of clusters and the availability of gold helpful passages for training.

free parameters (2)
  • Number of semantic clusters k = Not reported; selected based on Figure 4
    k controls the number of LoRA retrieval spaces trained in DPRM. Figure 4 tunes k against QA performance on the benchmarks, with no described held-out validation split.
  • Number of retrieved passages N per subquestion = Dataset-specific; not reported as a fixed value
    Figure 3 shows QA performance varying with N and the best N differs by dataset, indicating test-set-based selection rather than a fixed rule.
assumptions (5)
  • domain assumption Multi-hop questions can be decomposed into subquestions, and dependency links are correctly indicated by '#number#' markers emitted by QDM.
    Algorithm 1 and Section 3.3 use the marker as the sole dependency signal; if markers are wrong, SDOM rewrites the wrong subquestions.
  • domain assumption Semantic clusters and alignment patterns learned on 2WikiMQA training data transfer to HotpotQA and IIRC.
    Section 4.1 trains only on 2WikiMQA, and Section 3.3 maps out-of-domain subquestions to training-derived clusters; no domain adaptation is performed.
  • domain assumption A known 'helpful passage' is available for every training subquestion, and this question-passage alignment generalizes to unseen passages.
    The DPRM training objective in Eq. (3) requires a gold helpful passage p; in open-domain deployment such gold passages are not generally available.
  • ad hoc to paper ChatGPT-generated decomposition labels for QDM and SDOM are correct and diverse enough to serve as supervision.
    Section 4.3 states labels were generated by ChatGPT; no human verification or quality analysis is reported.
  • domain assumption E5-Mistral embeddings provide a meaningful basis for k-means clustering of subquestions before LoRA training.
    Section 3.2 clusters subquestions using base embeddings; poor clusters would assign subquestions to unrelated retrieval spaces.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimizing Question Semantic Space for Dynamic Retrieval-Augmented Multi-hop Question Answering." pith.science (2026). https://pith.science/paper/LANHTRVQ

@misc{pith2026250600491,
  author       = {Pith},
  title        = {Pith review of: Optimizing Question Semantic Space for Dynamic Retrieval-Augmented Multi-hop Question Answering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LANHTRVQ}},
  note         = {Machine review of arXiv:2506.00491}
}
read the original abstract

Retrieval-augmented generation (RAG) is usually integrated into large language models (LLMs) to mitigate hallucinations and knowledge obsolescence. Whereas,conventional one-step retrieve-and-read methods are insufficient for multi-hop question answering, facing challenges of retrieval semantic mismatching and the high cost in handling interdependent subquestions. In this paper, we propose Optimizing Question Semantic Space for Dynamic Retrieval-Augmented Multi-hop Question Answering (Q-DREAM). Q-DREAM consists of three key modules: (1) the Question Decomposition Module (QDM), which decomposes multi-hop questions into fine-grained subquestions; (2) the Subquestion Dependency Optimizer Module (SDOM), which models the interdependent relations of subquestions for better understanding; and (3) the Dynamic Passage Retrieval Module (DPRM), which aligns subquestions with relevant passages by optimizing the semantic embeddings. Experimental results across various benchmarks demonstrate that Q-DREAM significantly outperforms existing RAG methods, achieving state-of-the-art performance in both in-domain and out-of-domain settings. Notably, Q-DREAM also improves retrieval efficiency while maintaining high accuracy compared with recent baselines.

Figures

Figures reproduced from arXiv: 2506.00491 by the authors.

Figure 1
Figure 1. (a) Semantic Mismatching: A semantic prox [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall framework of Q-DREAM. model to generate the final answer. 3.2 Q-DREAM based Training In this section, we will introduce the training pro￾cess for the three modules of Q-DREAM. Question Decomposition Module (QDM) de￾composes Qori into subquestions via a model MQDM . For training the model, we optimize: max MQDM log P(Qsub | Qori; MQDM ) (1) where Qsub = {q1, . . . , qn}, and qi is a subques￾tion. Subquestion … view at source ↗
Figure 3
Figure 3. Impact of the number of retrieved passages. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

46 extracted references · 17 canonical work pages

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  2. [2]

    Yejin Bang, Samuel Cahyawijaya, Nayeon Lee, Wenliang Dai, Dan Su, Bryan Wilie, Holy Lovenia, Ziwei Ji, Tiezheng Yu, Willy Chung, et al. 2023. A multitask, multilingual, multimodal evaluation of chatgpt on reasoning, hallucination, and interactivity. arXiv preprint arXiv:2302.04023

    arXiv 2023

  3. [3]

    Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George Bm Van Den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, et al. 2022. Improving language models by retrieving from trillions of tokens. In International conference on machine learning, pages 2206--2240. PMLR

    2022

  4. [4]

    Kedi Chen, Qin Chen, Jie Zhou, Xinqi Tao, Bowen Ding, Jingwen Xie, Mingchen Xie, Peilong Li, and Zheng Feng. 2025. Enhancing uncertainty modeling with semantic graph for hallucination detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 23586--23594

    2025

  5. [5]

    Kedi Chen, Qin Chen, Jie Zhou, He Yishen, and Liang He. 2024. Diahalu: A dialogue-level hallucination evaluation benchmark for large language models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 9057--9079

    work page 2024

  6. [6]

    Bhuwan Dhingra, Jeremy R Cole, Julian Martin Eisenschlos, Daniel Gillick, Jacob Eisenstein, and William W Cohen. 2022. Time-aware language models as temporal knowledge bases. Transactions of the Association for Computational Linguistics, 10:257--273

    2022

  7. [7]

    James Ferguson, Matt Gardner, Hannaneh Hajishirzi, Tushar Khot, and Pradeep Dasigi. 2020. Iirc: A dataset of incomplete information reading comprehension questions. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1137--1147

    work page 2020

  8. [8]

    Nuno M Guerreiro, Duarte M Alves, Jonas Waldendorf, Barry Haddow, Alexandra Birch, Pierre Colombo, and Andr \'e FT Martins. 2023. Hallucinations in large multilingual translation models. Transactions of the Association for Computational Linguistics, 11:1500--1517

    work page 2023

Show all 46 references
  1. [9]

    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. [10]

    Xanh Ho, Anh-Khoa Duong Nguyen, Saku Sugawara, and Akiko Aizawa. 2020. Constructing a multi-hop qa dataset for comprehensive evaluation of reasoning steps. In Proceedings of the 28th International Conference on Computational Linguistics, pages 6609--6625

    2020

  3. [11]

    Minlie Huang, Xiaoyan Zhu, and Jianfeng Gao. 2020. Challenges in building intelligent open-domain dialog systems. ACM Transactions on Information Systems (TOIS), 38(3):1--32

    2020

  4. [12]

    Mohit Iyyer, Wen-tau Yih, and Ming-Wei Chang. 2017. Search-based neural structured learning for sequential question answering. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1821--1831

    2017

  5. [13]

    Gautier Izacard, Mathilde Caron, Lucas Hosseini, Sebastian Riedel, Piotr Bojanowski, Armand Joulin, and Edouard Grave. 2021. Unsupervised dense information retrieval with contrastive learning. arXiv preprint arXiv:2112.09118

    2021 arXiv

  6. [14]

    Gautier Izacard, Patrick Lewis, Maria Lomeli, Lucas Hosseini, Fabio Petroni, Timo Schick, Jane Dwivedi-Yu, Armand Joulin, Sebastian Riedel, and Edouard Grave. 2022. Few-shot learning with retrieval augmented language models. arXiv preprint arXiv:2208.03299

    2022 arXiv

  7. [15]

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al. 2023. Mistral 7b. arXiv preprint arXiv:2310.06825

    2023 arXiv

  8. [16]

    Ziheng Jiang, Haibin Lin, Yinmin Zhong, Qi Huang, Yangrui Chen, Zhi Zhang, Yanghua Peng, Xiang Li, Cong Xie, Shibiao Nong, et al. 2024. \ MegaScale \ : Scaling large language model training to more than 10,000 \ GPUs \ . In 21st USENIX Symposium on Networked Systems Design and...

    2024

  9. [17]

    Robertson

    K Sparck Jones, Steve Walker, and Stephen E. Robertson. 2000. A probabilistic model of information retrieval: development and comparative experiments: Part 2. Information processing & management, 36(6):809--840

    2000

  10. [19]

    Tushar Khot, Harsh Trivedi, Matthew Finlayson, Yao Fu, Kyle Richardson, Peter Clark, and Ashish Sabharwal. 2022 b . Decomposed prompting: A modular approach for solving complex tasks. arXiv preprint arXiv:2210.02406

    2022 arXiv

  11. [20]

    Jaehyung Kim, Jaehyun Nam, Sangwoo Mo, Jongjin Park, Sang-Woo Lee, Minjoon Seo, Jung-Woo Ha, and Jinwoo Shin. 2024. Sure: Summarizing retrievals using answer candidates for open-domain qa of llms. arXiv preprint arXiv:2404.13081

    2024 arXiv

  12. [21]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

    2014 arXiv

  13. [22]

    K Krishna and M Narasimha Murty. 1999. Genetic k-means algorithm. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 29(3):433--439

    1999

  14. [23]

    Angeliki Lazaridou, Elena Gribovskaya, Wojciech Stokowiec, and Nikolai Grigorev. 2022. Internet-augmented language models through few-shot prompting for open-domain question answering. arXiv preprint arXiv:2203.05115

    2022 arXiv

  15. [24]

    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

  16. [25]

    Yanming Liu, Xinyue Peng, Xuhong Zhang, Weihao Liu, Jianwei Yin, Jiannan Cao, and Tianyu Du. 2024. Ra-isf: Learning to answer and understand from retrieval augmentation via iterative self-feedback. arXiv preprint arXiv:2403.06840

    2024 arXiv

  17. [26]

    Xuan-Phi Nguyen, Shrey Pandit, Senthil Purushwalkam, Austin Xu, Hailin Chen, Yifei Ming, Zixuan Ke, Silvio Savarese, Caiming Xong, and Shafiq Joty. 2024. Sfr-rag: Towards contextually faithful llms. arXiv preprint arXiv:2409.09916

    2024 arXiv

  18. [27]

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 3...

    2022

  19. [28]

    Ofir Press, Muru Zhang, Sewon Min, Ludwig Schmidt, Noah A Smith, and Mike Lewis. 2023. Measuring and narrowing the compositionality gap in language models. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 5687--5711

    2023

  20. [29]

    Ori Ram, Yoav Levine, Itay Dalmedigos, Dor Muhlgay, Amnon Shashua, Kevin Leyton-Brown, and Yoav Shoham. 2023. In-context retrieval-augmented language models. Transactions of the Association for Computational Linguistics, 11:1316--1331

    2023

  21. [30]

    Sudha Rao and Hal Daum \'e III. 2019. Answer-based adversarial training for generating clarification questions. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and...

    2019

  22. [31]

    Weijia Shi, Sewon Min, Michihiro Yasunaga, Minjoon Seo, Rich James, Mike Lewis, Luke Zettlemoyer, and Wen-tau Yih. 2023. Replug: Retrieval-augmented black-box language models. arXiv preprint arXiv:2301.12652

    2023 arXiv

  23. [32]

    Alon Talmor and Jonathan Berant. 2018. The web as a knowledge-base for answering complex questions. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 641--651

    2018

  24. [33]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv e-prints, pages arXiv--2307

    2023

  25. [34]

    Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. 2023. Interleaving retrieval with chain-of-thought reasoning for knowledge-intensive multi-step questions. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Vo...

    2023

  26. [35]

    Boshi Wang, Xiang Deng, and Huan Sun. 2022. Shepherd pre-trained language models to develop a train of thought: An iterative prompting approach. arXiv preprint arXiv:2203.08383

    2022 arXiv

  27. [36]

    Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, and Furu Wei. 2023. Improving text embeddings with large language models. arXiv preprint arXiv:2401.00368

    2023 arXiv

  28. [37]

    Zhepei Wei, Wei-Lin Chen, and Yu Meng. 2025. Instruct RAG : Instructing retrieval-augmented generation via self-synthesized rationales. In The Thirteenth International Conference on Learning Representations

    2025

  29. [38]

    Tomer Wolfson, Mor Geva, Ankit Gupta, Matt Gardner, Yoav Goldberg, Daniel Deutch, and Jonathan Berant. 2020. Break it down: A question understanding benchmark. Transactions of the Association for Computational Linguistics, 8:183--198

    2020

  30. [39]

    Yuchen Xia, Jiho Kim, Yuhan Chen, Haojie Ye, Souvik Kundu, Cong Callie Hao, and Nishil Talati. 2024. Understanding the performance and estimating the cost of llm fine-tuning. In 2024 IEEE International Symposium on Workload Characterization (IISWC), pages 210--223. IEEE

    2024

  31. [40]

    Peng Xu, Wei Ping, Xianchao Wu, Chejian Xu, Zihan Liu, Mohammad Shoeybi, and Bryan Catanzaro. 2024. Chatqa 2: Bridging the gap to proprietary llms in long context and rag capabilities. arXiv preprint arXiv:2407.14482

    2024 arXiv

  32. [41]

    Jingfeng Yang, Haoming Jiang, Qingyu Yin, Danqing Zhang, Bing Yin, and Diyi Yang. 2022. Seqzero: Few-shot compositional semantic parsing with sequential prompts and zero-shot models. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 49--60

    2022

  33. [42]

    Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William Cohen, Ruslan Salakhutdinov, and Christopher D Manning. 2018. Hotpotqa: A dataset for diverse, explainable multi-hop question answering. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language ...

    2018

  34. [43]

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In International Conference on Learning Representations (ICLR)

    2023

  35. [44]

    Linhao Ye, Zhikai Lei, Jianghao Yin, Qin Chen, Jie Zhou, and Liang He. 2024. Boosting conversational question answering with fine-grained retrieval-augmentation and self-check. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Informa...

    2024

  36. [45]

    Yi-Fan Zhang, Huanyu Zhang, Haochen Tian, Chaoyou Fu, Shuangqing Zhang, Junfei Wu, Feng Li, Kun Wang, Qingsong Wen, Zhang Zhang, et al. 2024. Mme-realworld: Could your multimodal llm challenge high-resolution real-world scenarios that are difficult for humans? arXiv preprint a...

    2024 arXiv

  37. [46]

    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...

  38. [47]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.