Talk Before You Retrieve: Agent-Led Discussions for Better RAG in Medical QA

Hao Wang; Peijie Qiu; Rui Yin; Wenhui Zhu; Xiwen Chen; Xuanzhao Dong; Yalin Wang; Yi Su

REVIEW 4 major objections 6 minor 21 references

Talk Before You Retrieve: Agent-Led Discussions for Better RAG in Medical QA

T0 review · 4 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Adding an agent-led discussion stage before retrieval and a verification stage after it improves medical QA accuracy on four benchmarks, by up to 16.67 percentage points.

desk verdict A plausible plug-and-play RAG enhancement with a novel discuss-then-retrieve design, but the evaluation is too thin to establish that the gains come from better retrieval rather than the LLM feeding its own knowledge back through the summary. read the letter →

arxiv 2504.21252 v1 pith:7JJTNMNM submitted 2025-04-30 cs.CL

Xuanzhao Dong , Wenhui Zhu , Hao Wang , Xiwen Chen , Peijie Qiu , Rui Yin , Yi Su , Yalin Wang This is my paper · ORCID

classification cs.CL

keywords medicalquestionansweringretrieval-augmentedgenerationmulti-agentdiscussionLLMagentspost-retrievalverificationMedRAGBioASQPubQA

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

Discuss-RAG is a plug-and-play module that sits between a medical question and the usual retrieval step: before retrieving, a recruiter agent convenes a team of specialized medical agents who brainstorm the query over several turns, and a summarizer agent condenses their insights into a running summary that is appended to the original query. After retrieval, a decision-maker agent judges whether the returned snippets are relevant and coherent, and if not, triggers a chain-of-thought fallback retrieval. The paper reports that this two-stage design consistently outperforms the MedRAG baseline on MMLU-Med, MedQA-US, BioASQ, and PubMedQA, with the largest gains on BioASQ (58.61% to 75.28%) and PubMedQA (35.60% to 47.80%). Because both stages are training-free, the claim, if correct, means existing medical RAG pipelines can be made more accurate by adding discussion and verification rather than retraining the model.

What carries the argument

The load-bearing mechanism is the two-stage agent pipeline. Stage one is multi-turn discussion and summarization: at each turn $j$, expert agents $H_i$ each produce insight $I_i^j$, and a summarizer agent $C$ updates the running summary as $T^j := f_C(I_1^j,\ldots,I_n^j; T^{j-1}, Q)$; a verifier $V$ then distills the final summary into a verification-passed summary $D$ that is appended to $Q$ for retrieval. Stage two is post-retrieval verification: a decision-maker agent $U$ inspects the top-$k$ snippets $S_i$ and either accepts them for answer generation or triggers an alternative retrieval strategy. The paper's key design choice is that the expert agents are instructed not to answer the query, so the discussion stage is meant to build context for retrieval rather than produce the final answer.

What would settle it

Take a sample of BioASQ and PubMedQA questions, run only the discussion stage with the answers withheld, and check whether a human reader can answer the original question from the distilled summary $D$ alone; if $D$ answers a substantial fraction of questions, the no-answer instruction is not honored and the retrieval comparison is confounded. Alternatively, replace $D$ with a topic-preserving paraphrase stripped of answer-specific terms and measure whether the accuracy gains disappear.

Watch

Extended reading notes

Core claim

The core discovery is that the quality of retrieved snippets, not the generator alone, is the bottleneck in medical RAG, and that agent-led context construction changes what the retriever returns. In Discuss-RAG, the distilled verification-passed summary $D$, produced by a team of expert agents and a summarizer, is concatenated with the original query $Q$ and fed to the retriever; this yields snippets that are more topically relevant and better ordered, as illustrated in the MedQA-US oxygen-saturation example. The decision-maker agent then gates the snippets before answer generation, filtering out irrelevant or misleading evidence and using a chain-of-thought prompt as a fallback when the retrieved set looks poor. On the four benchmarks, the paper reports consistent accuracy improvements over MedRAG, including large gains on BioASQ and PubMedQA, and attributes them to better snippet relevance and coherence rather than to modifications of the answer generator.

Load-bearing premise

The method assumes the expert agents, even when told not to answer the query, produce discussion summaries that improve retrieval without already containing the correct answer; if the summary leaks the answer, the reported retrieval gains are partly the model retrieving what it already wrote.

Editorial extensions

If this is right

If the reported gains are real, any training-free medical RAG pipeline can likely be improved by inserting a discussion-based query enrichment step and a snippet-verification step, independent of the retriever and generator used.
The larger gains on BioASQ and PubMedQA, where baseline accuracy is lower, suggest the module is most valuable when retrieval noise is high or corpus coverage is weak.
The decision-maker's fallback mechanism implies that retrieval failures can be detected before answer generation, so a system could selectively apply more expensive retrieval strategies only when needed.
Because the module changes only the retriever's input and the selection of snippets, it should combine additively with better retrievers, better LLMs, or other RAG enhancements.

Reading between the lines

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

An editorial extension: the same discuss-before-retrieve, verify-after-retrieve pattern could transfer to other high-stakes QA domains, such as legal or scientific literature, where surface-similarity retrieval is also a weak signal.
A testable extension: replace the multi-agent discussion with a single agent writing a background summary under the same no-answer instruction; if accuracy stays at the Discuss-RAG level, the team discussion is unnecessary, and if it drops, the multi-perspective interaction is the active ingredient.
A cost-saving extension: because the authors note the framework adds computational overhead, distilled summaries could be cached or reused for similar queries rather than regenerated from scratch each time.

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.

Desk Editor's Note

read the letter

You should know that Discuss-RAG is a genuinely new combination: a multi-agent discussion and summarization step before retrieval, plus a post-retrieval verifier. The reported gains on four medical QA benchmarks are large, and if they hold, this is a useful training-free plug-in for applied systems. But the evidence is thin, and I don't think the central mechanism is yet established.

What's new: prior work does post-hoc verification (chain-of-verification) or multi-agent decision-making (MDAgents), but nobody has used agent-led discussion to construct a retrieval query. The idea is sensible—enrich the query with domain background before hitting the retriever—and the qualitative examples show the retrieved snippets arguably become more relevant. The method is modular, and the code is public, which helps.

The soft spots are real. The evaluation compares only against MedRAG, with no repeated runs. GPT-3.5 is stochastic; a 16.67-point gain on BioASQ could be partly run-to-run volatility. There's no ablation of either module, so you can't tell how much the discussion module contributes versus the verifier. And the leakage question is legitimately unresolved. The agents are instructed not to answer the query, but the paper never checks whether the distilled summary D contains answer-relevant content. The example in Fig. 4 is suggestive but not damning: D mentions 'lowest oxygen content,' which is the query phrase, not the answer 'coronary sinus.' Still, a simple control—like comparing against a summary that just paraphrases the question without domain expertise—would settle it. The limitation section acknowledges compute and limited interaction, but oddly not this verification gap.

Who is this for? Anyone working on RAG query expansion, agentic retrieval, or medical QA will want to know about this design. It's a useful empirical kickoff, not a settled result. I'd send it to peer review; a good referee can push for the missing ablations and leakage controls, and the idea deserves a fair shot.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes Discuss-RAG, a plug-and-play module for medical retrieval-augmented generation (RAG) that first orchestrates a team of medical expert agents to discuss the query and produce a distilled summary D, which is then concatenated with the original query for retrieval; a separate decision-maker agent post-processes the retrieved snippets before answer generation. The authors evaluate on four medical QA benchmarks (MMLU-Med, MedQA-US, BioASQ, PubMedQA) against the MedRAG baseline and report consistent accuracy improvements, with the largest gains on BioASQ (+16.67%) and PubMedQA (+12.20%). The paper also includes a motivating pilot study and a qualitative example. The code is publicly available.

Significance. The proposed idea of using agent-led discussion to enrich retrieval queries in medical RAG is intuitive and practically relevant, and the modular design means it could be dropped into existing training-free pipelines. The paper ships code and reports positive results on four public benchmarks. However, the current evaluation is too thin to establish the central claim: the comparison is limited to a single baseline, there are no error bars or significance tests, and the two proposed components are not ablated. Most importantly, the paper does not rule out the possibility that the distilled summary D leaks answer-relevant content back into the retrieval step, in which case the gains would be driven by the LLM rediscovering its own parametric knowledge rather than by improved retrieval relevance. If these concerns are addressed, the contribution would be a useful empirical result; at present the evidence is suggestive rather than conclusive.

major comments (4)

[Section 3 and Figure 4(B)] The instruction that agents 'not answer the original query' is asserted but not verified. The qualitative example in Figure 4(B) shows the distilled summary D stating that 'oxygen extraction by tissues plays a significant role in determining the area with the lowest oxygen content' and highlighting 'lowest oxygen content at both time points,' which closely track the reasoning steps needed to select 'coronary sinus.' Because D is concatenated with Q and fed to the retriever, answer-related content in D can steer retrieval toward documents that confirm the correct answer. The paper provides no control for this leakage: no measurement of answer-term overlap between D and the gold answer, no ablation replacing D with a generic medical summary, and no comparison to a summary generated from a query with answer options removed. Without such a control, the reported gains cannot be attributed to improved retrieval relevance rather than to the LLM's own output being recycled through the pipeline.
[Section 4, Table 1] The evaluation reports a single run for each benchmark and no variance or significance measures. Since GPT-3.5 (gpt-3.5-turbo-0125) is stochastic and the method involves multiple LLM calls, the reported improvements (e.g., +16.67% on BioASQ, +12.20% on PubMedQA) may not be stable. The authors should report results over multiple runs (at least three to five) with mean and standard deviation, and test whether the differences against MedRAG are statistically significant (e.g., with a paired test or bootstrap). This is particularly important because gains vary considerably across datasets.
[Section 4, Table 1] The paper does not include ablations of its two proposed modules: the multi-turn discussion/summarization and the post-retrieval verification. Without ablations, one cannot determine which component drives the improvement, or whether the verifier alone (without the discussion) would yield similar gains. Additionally, the paper compares only to MedRAG; a stronger evaluation would include at least one more recent RAG baseline (e.g., self-RAG or chain-of-verification RAG) to place the improvement in context.
[Sections 3 and 4] Several implementation details needed for reproducibility are missing: the number of expert agents n, the maximum discussion rounds m, the number of retrieved chunks k, and the LLM temperature are not reported. The description of the post-retrieval verification is also underspecified: when the decision-maker U returns a negative judgment, the paper says an 'alternative retrieval strategy (e.g., a CoT-based prompt)' is triggered, but it does not describe what this fallback is, how often it fires, or whether it changes the retrieved set. These details are necessary for others to reproduce or build on the method.

minor comments (6)

[Section 2] The text says 'As shown in Fig. 2(A), when the corpus is fixed (i.e., textbooks (Jin et al., 2021)), varying the number of retrieved documents k results in fluctuating accuracy across six medical subjects.' The text could be clearer about whether the data are aggregated over the six subjects or shown separately, and whether the setup matches the later experiments.
[Section 4] The sentence 'MedRAG (Xiong et al., 2024).' is incomplete; it should be integrated into a full sentence describing the baseline and its role.
[Section 1] The contribution list says 'comprehensive experiments comparing Discuss-RAG with standard RAG systems,' but the evaluation compares only to MedRAG. Please revise the wording to match the actual scope of the experiments.
[Figure 4 and surrounding text] The typo 'Disscuss-RAG' appears in the text near Figure 4; correct it to 'Discuss-RAG.'
[Equation (1)] The notation T^j := f_C(I^j_1, ..., I^j_n ; T^{j-1}, Q) is confusing because the superscripts on I and T are inconsistent (the paper writes I^j_i but also I_j^i in the text). Please define the indices clearly and explain what f_C takes as input.
[Section 3] The recruiter agent R is introduced but its role is not described in detail; the text should clarify how R assembles the team, whether it is a separate LLM call, and how the team composition is chosen for each query.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the benchmarks are external, no parameter fitting or author-self-citation is load-bearing, and the unverified no-answer prompt is a control concern rather than a demonstrated derivation circle.

full rationale

The paper's contribution is an empirical pipeline: given a medical query Q, an agent team produces a distilled summary D that is used together with Q to retrieve from a fixed textbook corpus, and a decision-maker filters snippets before final answer generation. The reported accuracy numbers on MMLU-Med, MedQA-US, BioASQ, and PubMedQA are external benchmark scores, not quantities derived from the method's own inputs by construction. There is no fitted parameter that is later renamed as a prediction, no equation equates the reported gains with a prior fit, and the MedRAG baseline is an external system cited from other authors, not prior work by the present authors. The strongest circularity-adjacent concern is that the agents are only 'explicitly instructed not to answer the original query or infer a final conclusion,' and the paper does not measure whether D leaks answer-relevant content. The qualitative example in Fig. 4(B) shows a summary that highlights 'oxygen extraction by tissues,' which could conceivably steer retrieval toward snippets supporting the correct answer. However, the paper does not exhibit the required specific reduction: there is no demonstration that D equals or encodes the answer, no control ablation comparing D against a generic medical summary, and no fitted quantity being fed back as a prediction. A missing compliance check is a legitimate experimental weakness and a falsifiability concern, but it is not circularity unless the paper shows that the improvement is forced by construction. Under the hard rule that circularity requires quoted evidence of a specific reduction, the honest finding is no significant circularity; the answer-leakage worry belongs in the correctness-risk discussion, not the circularity score.

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

The method's central claim rests on design choices (n, m, k, temperature) and on the untested assumption that expert-agent summaries do not leak answers. No parameters are fitted to the test sets, but these hyperparameters are not reported, hurting reproducibility.

free parameters (4)

number of expert agents n
Not specified in the paper; the example uses three agents (Sec. 3, Fig. 3). The choice likely affects summary richness and retrieval quality.
maximum discussion rounds m
The loop terminates at m rounds or when agents decline; m is not reported. Different m changes the distilled summary.
number of retrieved chunks k
The paper varies k in preliminary experiments (Fig. 2A) but does not state the k used for the main results, inheriting the MedRAG default.
LLM temperature
Not reported; GPT-3.5 outputs are stochastic, so temperature affects the accuracy numbers and reproducibility.

assumptions (3)

domain assumption Agents provide relevant domain background without answering the query
The no-answer instruction is only a prompt; the paper does not verify compliance or measure leakage (Sec. 3).
domain assumption The fixed corpus of medical textbooks contains the knowledge needed for all four benchmarks
All experiments use the same textbook corpus (Jin et al., 2021); if a question is not covered, retrieval cannot help.
domain assumption MedCPT is a suitable retriever for medical snippets
Borrowed from MedRAG without independent validation; the agent summary must be compatible with this retriever.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Talk Before You Retrieve: Agent-Led Discussions for Better RAG in Medical QA." pith.science (2026). https://pith.science/paper/7JJTNMNM

@misc{pith2026250421252,
  author       = {Pith},
  title        = {Pith review of: Talk Before You Retrieve: Agent-Led Discussions for Better RAG in Medical QA},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7JJTNMNM}},
  note         = {Machine review of arXiv:2504.21252}
}

read the original abstract

Medical question answering (QA) is a reasoning-intensive task that remains challenging for large language models (LLMs) due to hallucinations and outdated domain knowledge. Retrieval-Augmented Generation (RAG) provides a promising post-training solution by leveraging external knowledge. However, existing medical RAG systems suffer from two key limitations: (1) a lack of modeling for human-like reasoning behaviors during information retrieval, and (2) reliance on suboptimal medical corpora, which often results in the retrieval of irrelevant or noisy snippets. To overcome these challenges, we propose Discuss-RAG, a plug-and-play module designed to enhance the medical QA RAG system through collaborative agent-based reasoning. Our method introduces a summarizer agent that orchestrates a team of medical experts to emulate multi-turn brainstorming, thereby improving the relevance of retrieved content. Additionally, a decision-making agent evaluates the retrieved snippets before their final integration. Experimental results on four benchmark medical QA datasets show that Discuss-RAG consistently outperforms MedRAG, especially significantly improving answer accuracy by up to 16.67% on BioASQ and 12.20% on PubMedQA. The code is available at: https://github.com/LLM-VLM-GSL/Discuss-RAG.

Figures

Figures reproduced from arXiv: 2504.21252 by the authors.

**Figure 2.** Preliminary experiments on the MMLU-Med benchmark. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Illustration of the Discuss-RAG pipeline. (A). depicts the multi-turn brainstorming and summarization process. (B). presents the agent-led post-retrieval verification module. A medical query, the corresponding snippets, and the LLM’s generated answer are used for illustration. Further details are provided in Sec 3. To further examine the limitations of hard similarity-based retrieval, we conducted an exploratory ex… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Example from the MedQA-US benchmark comparing MedRAG [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 4 canonical work pages

[1]

Scott Barnett, Stefanus Kurniawan, Srikanth Thudumu, Zach Brannelly, and Mohamed Abdelrazek. 2024. Seven failure points when engineering a retrieval augmented generation system. In Proceedings of the IEEE/ACM 3rd International Conference on AI Engineering-Software Engineering for AI, pages 194--199

work page 2024
[2]

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
[3]

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
[4]

Bolei He, Nuo Chen, Xinran He, Lingyong Yan, Zhenkai Wei, Jinchang Luo, and Zhen-Hua Ling. 2024. Retrieving, rethinking and revising: The chain-of-verification can improve retrieval augmented generation. arXiv preprint arXiv:2410.05801

arXiv 2024
[5]

Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2020. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300

arXiv 2020
[6]

Gautier Izacard and Edouard Grave. 2020. Leveraging passage retrieval with generative models for open domain question answering. arXiv preprint arXiv:2007.01282

arXiv 2020
[7]

Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM computing surveys, 55(12):1--38

2023
[8]

Di Jin, Eileen Pan, Nassim Oufattole, Wei-Hung Weng, Hanyi Fang, and Peter Szolovits. 2021. What disease does this patient have? a large-scale open domain question answering dataset from medical exams. Applied Sciences, 11(14):6421

2021

Show all 21 references

[9]

Qiao Jin, Bhuwan Dhingra, Zhengping Liu, William W Cohen, and Xinghua Lu. 2019. Pubmedqa: A dataset for biomedical research question answering. arXiv preprint arXiv:1909.06146

2019 arXiv
[10]

Qiao Jin, Won Kim, Qingyu Chen, Donald C Comeau, Lana Yeganova, W John Wilbur, and Zhiyong Lu. 2023. Medcpt: Contrastive pre-trained transformers with large-scale pubmed search logs for zero-shot biomedical information retrieval. Bioinformatics, 39(11):btad651

2023
[11]

YuHe Ke, Liyuan Jin, Kabilan Elangovan, Hairil Rizal Abdullah, Nan Liu, Alex Tiong Heng Sia, Chai Rick Soh, Joshua Yi Min Tung, Jasmine Chiat Ling Ong, and Daniel Shu Wei Ting. 2024. Development and testing of retrieval augmented generation in large language models--a case stu...

2024 arXiv
[12]

Yubin Kim, Chanwoo Park, Hyewon Jeong, Yik Siu Chan, Xuhai Xu, Daniel McDuff, Hyeonhoon Lee, Marzyeh Ghassemi, Cynthia Breazeal, Hae Park, et al. 2024. Mdagents: An adaptive collaboration of llms for medical decision-making. Advances in Neural Information Processing Systems, 3...

2024
[13]

Harsha Nori, Nicholas King, Scott Mayer McKinney, Dean Carignan, and Eric Horvitz. 2023. Capabilities of gpt-4 on medical challenge problems. arXiv preprint arXiv:2303.13375

2023 arXiv
[14]

OpenAI. 2024. Gpt-3.5 turbo. https://platform.openai.com/docs/models/gpt-3-5-turbo. Accessed: 2025-04-27

2024
[15]

Karan Singhal, Shekoofeh Azizi, Tao Tu, S Sara Mahdavi, Jason Wei, Hyung Won Chung, Nathan Scales, Ajay Tanwani, Heather Cole-Lewis, Stephen Pfohl, et al. 2023. Large language models encode clinical knowledge. Nature, 620(7972):172--180

2023
[16]

George Tsatsaronis, Georgios Balikas, Prodromos Malakasiotis, Ioannis Partalas, Matthias Zschunke, Michael R Alvers, Dirk Weissenborn, Anastasia Krithara, Sergios Petridis, Dimitris Polychronopoulos, et al. 2015. An overview of the bioasq large-scale biomedical semantic indexi...

2015
[17]

Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824--24837

2022
[18]

Guangzhi Xiong, Qiao Jin, Zhiyong Lu, and Aidong Zhang. 2024. Benchmarking retrieval-augmented generation for medicine. In Findings of the Association for Computational Linguistics ACL 2024, pages 6233--6251

2024
[19]

Zihan Zhang, Meng Fang, Ling Chen, Mohammad-Reza Namazi-Rad, and Jun Wang. 2023. How do large language models capture the ever-changing world knowledge? a review of recent advances. arXiv preprint arXiv:2310.07343

2023 arXiv
[20]

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...
[21]

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 16, 2026 · model on record in the stance chip above.

[1] [1]

Scott Barnett, Stefanus Kurniawan, Srikanth Thudumu, Zach Brannelly, and Mohamed Abdelrazek. 2024. Seven failure points when engineering a retrieval augmented generation system. In Proceedings of the IEEE/ACM 3rd International Conference on AI Engineering-Software Engineering for AI, pages 194--199

work page 2024

[2] [2]

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

[3] [3]

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

[4] [4]

Bolei He, Nuo Chen, Xinran He, Lingyong Yan, Zhenkai Wei, Jinchang Luo, and Zhen-Hua Ling. 2024. Retrieving, rethinking and revising: The chain-of-verification can improve retrieval augmented generation. arXiv preprint arXiv:2410.05801

arXiv 2024

[5] [5]

Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2020. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300

arXiv 2020

[6] [6]

Gautier Izacard and Edouard Grave. 2020. Leveraging passage retrieval with generative models for open domain question answering. arXiv preprint arXiv:2007.01282

arXiv 2020

[7] [7]

Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM computing surveys, 55(12):1--38

2023

[8] [8]

Di Jin, Eileen Pan, Nassim Oufattole, Wei-Hung Weng, Hanyi Fang, and Peter Szolovits. 2021. What disease does this patient have? a large-scale open domain question answering dataset from medical exams. Applied Sciences, 11(14):6421

2021