Large Language Model Can Be a Foundation for Hidden Rationale-Based Retrieval

Feixiang Guo; Hongyang Lei; Luo Ji; Ningyuan Xi; Peng Yu; Qingqing Gu; Teng Chen; Xiaoyu Wang; Yihong Wang; Yong Chen

REVIEW 4 major objections 5 minor 34 references

Large Language Model Can Be a Foundation for Hidden Rationale-Based Retrieval

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an instruction-tuned LLM used as a cross-encoder, prompted with the binary-choice question 'Can Q be appropriately responded with D?' and scored by relative next-token probabilities, outperforms existing bi-encoder…

desk verdict Useful new retrieval task and strong results, but Eq. (2) as written inverts the ranking and must be fixed before the numbers can be trusted. read the letter →

arxiv 2412.16615 v2 pith:YLM5PY5Q submitted 2024-12-21 cs.IR cs.CLcs.LG

Luo Ji , Feixiang Guo , Teng Chen , Qingqing Gu , Xiaoyu Wang , Ningyuan Xi , Yihong Wang , Peng Yu

show 4 more authors

Yue Zhao Hongyang Lei Zhonglin Jiang Yong Chen

This is my paper · ORCID

classification cs.IRcs.CLcs.LG

keywords hiddenrationaleretrievalgenerativecross-encoderLLMdirectpreferenceoptimizationbinary-choicepromptingemotionalsupportconversationprefixcachingretrieval-augmentedgeneration

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

The paper proposes and tests the claim that an instruction-tuned LLM used as a cross-encoder, trained with a purely generative loss, can serve as a foundation for hidden-rationale retrieval—tasks where the query and the correct document are not semantically similar but are linked by reasoning chains, logic, or experience. The method, called LaHoRe, converts retrieval into a binary-choice generation problem by asking 'Can Q be appropriately responded with D?' and scoring each candidate by the relative next-token probability of versus . On four emotional-support and intent-retrieval datasets, LaHoRe beats all prior bi-encoder and cross-encoder baselines in zero-shot evaluation, and improves further with supervised fine-tuning or direct preference optimization. The authors also show that putting the document before the query allows offline caching of the document's key-value states, reducing per-query response time by roughly 60% without harming accuracy.

What carries the argument

The load-bearing object is the binary-choice retrieval instruction, which turns retrieval into a text-generation task: the appended prompt {Ir}='Can Q be appropriately responded with D?' together with {Ibc}='If you think the answer is true, choose <T>; otherwise choose <F>.' The relevance score is read out as the relative confidence s_rel = s(<T>)/(s(<T>)+s(<F>)) from the LLM's next-token log-probabilities. The supporting machinery is the D⇒Q input ordering, which lets the document's key-value cache be computed offline and reused across queries, making the cross-encoder retriever practical while preserving most of the zero-shot accuracy.

What would settle it

Re-annotate the ESconv and PsyQA test queries with all acceptable reply strategies rather than a single one, then compare LaHoRe against the baselines under that multi-label evaluation; if the gap shrinks or reverses, the reported superiority is an artifact of the single-label setup.

Watch

Extended reading notes

Core claim

The central discovery, stated on the paper's own terms, is that discriminative retrieval can be reformulated as a generative preference judgment. Instead of computing a similarity score or a contrastive embedding, the retriever feeds the LLM the concatenation 'document: {D}\n query: {Q}\n' plus a special instruction asking whether Q can be appropriately responded with D and saying 'If you think the answer is true, choose <T>; otherwise choose <F>.' The relevance score is s_rel = s(<T>)/(s(<T>)+s(<F>)), where s(A) is the log next-token probability of answer A. With this setup, a 7B instruction-tuned LLM outperforms all compared bi-encoder and cross-encoder retrievers on hidden-rationale benchmarks, both zero-shot and after fine-tuning, and DPO fine-tuning yields the strongest results. The paper further claims the same framework can be made efficient by ordering the document before the query and caching its prefix.

Load-bearing premise

Each user query is assumed to have exactly one correct retrieval target, because every dataset labels each query with a single strategy or intent; if those labels are noisy or the strategies overlap, the reported Recall and p-MRR numbers stop being a reliable measure of retrieval quality.

Editorial extensions

If this is right

On all four datasets (ChatIntent, ChatStrategy, ESconv, PsyQA), LaHoRe attains the best or second-best zero-shot score on every metric and the best fine-tuned score on every metric among 7B-parameter baselines.
DPO fine-tuning beats SFT fine-tuning on every dataset, indicating that pairwise preference optimization aligns the model better with hidden-rationale judgments than supervised labeling.
Using every mismatched document as a negative (a 1:all ratio) gives the best accuracy, and the benefit grows with the size of the candidate pool.
Swapping the order to D⇒Q cuts response time from roughly 47.5 ms to 19.3 ms per query in the reported tests, with accuracy roughly unchanged.
In an end-to-end RAG setting, LaHoRe-retrieved strategies produce responses that human judges prefer over raw LLM output 90% of the time.

Reading between the lines

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

If the advantage transfers beyond emotional support, the same yes/no scoring could serve as a general plug-in retriever for other non-factual tasks, such as choosing an embodied subgoal from a candidate list generated by an LLM planner.
Because the method only needs next-token probabilities, any instruction-tuned LLM could be used as the retriever without architectural changes; a natural extension is to test how retrieval quality scales with model size and instruction-tuning quality.
The single-label evaluation may be hiding ambiguity: emotional-support strategies overlap, so re-annotating test queries with all acceptable strategies could show whether LaHoRe's ranking stays plausible when more than one answer is legitimately correct.
The DPO-over-SFT result suggests that preference data is a stronger training signal than binary labels for retrieval; a direct comparison of DPO against a contrastive-loss variant with identical negative sampling would isolate what the generative formulation itself contributes.

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

LaHoRe is worth a serious look. The paper names a real gap—retrieval where query and document are linked by reasoning rather than surface similarity—and shows a 7B instruction-tuned LLM, used as a generative cross-encoder with a binary-choice prompt, beats established bi- and cross-encoders on emotional-support strategy and intent retrieval. The D⇒Q prompt order with KV-cache reuse is a practical contribution, and the SFT/DPO gains are large. If the code and data are actually released, that adds real value.

The load-bearing problem is Eq. (2). It defines the relevance score as s()/(s()+s()) where s(A)=log P(A). Because both logs are negative and the denominator is log(p(1-p)), this ratio decreases as P() rises. Taken literally, the method ranks documents with lower probability of higher. That would invert every zero-shot result in Table 2. The likely intended form is a softmax over the log probabilities—i.e., exp(s())/(exp(s())+exp(s())) = P(). As written, the paper's stated scoring rule cannot reproduce the reported scores. This is an internal inconsistency, not a disagreement about methodology.

The other issues are secondary but real. There are no error bars or multiple seeds, so the margin over baselines is hard to judge. The sensitivity analysis for the positive-negative ratio uses what looks like the test sets, so the 1:all choice may be selected on the same data used for the final numbers. DPO pair construction is underspecified (how are negatives paired? what reference model? what beta?). The single-label evaluation on ESconv and PsyQA is also questionable because emotional-support strategies are not mutually exclusive; a response can validly combine several strategies, so Recall/p-MRR against one annotated label is a noisy measure. Two of the four datasets are proprietary, though the repo may mitigate that.

My read: the idea is solid and the authors are onto something, but the paper as written cannot be reproduced from the equations. A referee should not desk-reject; they should ask for a corrected scoring rule, re-run or explicit confirmation, error bars, validation-based hyperparameter selection, and DPO details. If Eq. (2) is just a missing exp, the paper is a likely accept after minor revision.

Referee Report

4 major / 5 minor

Summary. The paper introduces LaHoRe, a cross-encoder LLM retriever designed for "hidden rationale retrieval," where query and document are not semantically similar but related by reasoning, logic, or empirical knowledge. The method appends a binary-choice instruction ("Can Q be appropriately responded with D?") to the query--document pair, converts retrieval into a next-token generation problem, and scores relevance from the relative next-token probabilities of the two choice tokens. The authors evaluate zero-shot and SFT/DPO fine-tuned versions on four datasets constructed from ESconv, PsyQA, and a proprietary Daily Chat corpus, reporting that LaHoRe outperforms existing bi-encoder and cross-encoder baselines on Recall and p-MRR. They also propose a query/document reordering that enables KV-cache reuse for faster inference, and report an end-to-end RAG win-rate of 90% over raw LLM generation.

Significance. If the empirical claims hold, the paper makes a meaningful contribution by broadening retrieval beyond semantic similarity, demonstrating that a purely generative loss can serve as an alternative to contrastive learning for relevance scoring, and providing a practical inference-speed optimization via prompt caching. The release of code, models, and datasets is a concrete strength, and the evaluation design is not circular: zero-shot results are obtained with an unmodified instruction-tuned Qwen2 and externally defined strategy/intent labels. However, the significance is conditional on resolving a formal inconsistency in the scoring rule, clarifying the evaluation protocol with respect to validation and seed variance, and specifying the DPO data construction.

major comments (4)

[Section 2.2, Eq. (2)] Equation (2) defines the relevance score as s_rel = s(<T>) / (s(<T>) + s(<F>)), where s(A) = log P(A|input). When the binary-choice prompt is used, P(<T>) + P(<F>) = 1, so both log probabilities are negative and s_rel is decreasing in P(<T>). For example, P(<T>) = 0.6 gives s_rel ≈ 0.36, while P(<T>) = 0.2 gives s_rel ≈ 0.88. Taken literally, the stated rule would rank documents with lower probability of '<T>' as more relevant, which is the opposite of the reported zero-shot results in Table 2. Either an exp() is missing (the intended softmax over log probabilities) or the implemented scoring rule is different from the one stated. Because every zero-shot and fine-tuned result depends on this score, the central claim cannot be reproduced from the text as written. This must be corrected and clarified.
[Sections 3.2 and 3.3] The paper does not describe a validation split or a model-selection protocol. The train-test split is stated as 9:1, and the sensitivity analysis over positive-negative ratios (Fig. 3) appears to be conducted on the same test sets used for the fine-tuning results in Table 3. If the 1:all ratio and other hyperparameters were selected after observing test-set performance, the reported fine-tuning comparisons are optimistic and not a fair evaluation. Please specify the validation procedure, report results with multiple seeds and error bars, and, if necessary, re-run model selection on a held-out split.
[Section 2.3] The construction of DPO preference pairs is unspecified. The text explains that each query is combined with mismatched documents to create negative pairs and that labels are '<T>' for positive and '<F>' for negative pairs, which is sufficient for SFT, but DPO requires chosen/rejected pairs. The paper does not state how these pairs are formed, whether all negatives are used, or how ties are handled. This missing detail prevents reproduction of the LaHoRe-DPO results.
[Section 3.1] The evaluation assumes that each user utterance has exactly one correct retrieval target. In emotional support conversations, strategies such as 'Reflection of feelings' and 'Affirmation and Reassurance' are not mutually exclusive, and a single response may legitimately combine several strategies. If the single-label annotations are noisy or ambiguous, the reported Recall and p-MRR values are less reliable as measures of retrieval quality. Please report label-agreement statistics or discuss how ambiguity affects the comparison.

minor comments (5)

[Abstract and Section 1] The model is named 'LaHoRe' in the title and method sections, but the abstract and one introductory sentence call it 'RaHoRe.' Please unify the terminology.
[Section 2.2] The symbol A in Eq. (2) is used before being defined; please state explicitly that A is the model's next-token choice, i.e., '<T>' or '<F>'.
[Section 3.3] The sensitivity paragraph says '1.5' where the figure and context indicate '1:5', and 'the formal choice' should read 'the former choice.' Additionally, 'positive outnumbered by negative samples' should be rephrased.
[Table 1 and Table 4] There are spacing and formatting artifacts in the table captions ('T able 1') and in the 'RT↓' column; please clean these up and specify the hardware and measurement protocol for response time.
[Section 2.2] The claim about 'prefix decoding or cache-sharing technology' should cite the specific mechanism used (e.g., prompt caching or KV-cache reuse) and state whether the reported RT improvements come from the reordering alone or from an additional caching implementation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the central zero-shot claim is independently grounded, though Eq. (2) is internally inconsistent.

full rationale

LaHoRe's central claim does not reduce to its inputs. The zero-shot evaluation uses an unmodified instruction-tuned Qwen2-7B with a fixed binary-choice prompt and scores documents by the model's next-token probabilities; the ground-truth strategy/intent labels come from external datasets (ESConv, PsyQA, Daily Chat), so the ranking is not defined in terms of the target labels. Fine-tuned SFT/DPO results are standard supervised evaluations on held-out test splits and do not rename fitted parameters as predictions. The 1:all negative-sampling choice in Section 2.3 is a training-data construction, and the sensitivity analysis in Section 3.3 is an empirical observation, not a derivation; although Fig. 3 appears to evaluate on test sets, this is a potential data-selection bias rather than a circular argument. The paper contains no load-bearing self-citation chain: the R-tuning citation [26] is only for the binary-choice prompt format and is not the basis of the retrieval claim. One serious non-circular defect is Eq. (2): if the binary prompt enforces P(<T>) + P(<F>) = 1, the ratio log P(<T>)/(log P(<T>) + log P(<F>)) is decreasing in P(<T>), so a literal reading would invert the ranking; this makes the reported results irreproducible from the stated scoring rule and likely reflects a missing exp(), but it is a correctness/reproducibility problem, not a circularity. Overall the derivation chain is therefore self-contained, with no step where a prediction is equivalent to its inputs by construction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central empirical claim rests on the validity of the single-label retrieval datasets and on the assumption that next-token probabilities of the binary tokens encode relevance. No new physical entities are introduced. The free parameters are standard training hyperparameters plus the 1:all negative sampling ratio, which was chosen after inspecting sensitivity results.

free parameters (5)

learning_rate = 1e-6
Set in Section 3.2 for all fine-tuning runs; no sweep or sensitivity report is given.
epochs = 3
Set in Section 3.2; no ablation is reported.
batch_size = 512 (SFT), 128 (DPO)
Set in Section 3.2; no ablation is reported.
window_length = 2048
Set in Section 3.2; no sensitivity analysis is reported.
positive_negative_ratio = 1:all
Chosen after comparing 1:1, 1:3, 1:5, 1:10, and 1:all in Fig. 3; appears selected on test data, introducing selection bias.

assumptions (4)

domain assumption A single correct retrieval target exists for each query in all four datasets.
Section 3.1 builds retrieval datasets from annotated strategy and intent labels; if multiple strategies are plausible, measured Recall and p-MRR are not reliable.
domain assumption Next-token log probabilities of <T> and <F> reflect true relevance for ranking.
Eq. (2) defines the relevance score from these probabilities, but the paper provides no calibration analysis or failure cases.
domain assumption An instruction-tuned LLM answers consistently with <T> or <F> under the binary-choice prompt.
Section 2.1 assumes the model constrains its answer to the two tokens; no parsing or refusal-rate analysis is reported.
standard math Prefix caching of document encodings is valid for cross-encoder scoring with a causal LLM.
Section 2.2 uses causal decoding so D-before-Q allows KV cache reuse; this is a standard property of causal transformer attention.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Large Language Model Can Be a Foundation for Hidden Rationale-Based Retrieval." pith.science (2026). https://pith.science/paper/YLM5PY5Q

@misc{pith2026241216615,
  author       = {Pith},
  title        = {Pith review of: Large Language Model Can Be a Foundation for Hidden Rationale-Based Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YLM5PY5Q}},
  note         = {Machine review of arXiv:2412.16615}
}

read the original abstract

Despite the recent advancement in Retrieval-Augmented Generation (RAG) systems, most retrieval methodologies are often developed for factual retrieval, which assumes query and positive documents are semantically similar. In this paper, we instead propose and study a more challenging type of retrieval task, called hidden rationale retrieval, in which query and document are not similar but can be inferred by reasoning chains, logic relationships, or empirical experiences. To address such problems, an instruction-tuned Large language model (LLM) with a cross-encoder architecture could be a reasonable choice. To further strengthen pioneering LLM-based retrievers, we design a special instruction that transforms the retrieval task into a generative task by prompting LLM to answer a binary-choice question. The model can be fine-tuned with direct preference optimization (DPO). The framework is also optimized for computational efficiency with no performance degradation. We name this retrieval framework by RaHoRe and verify its zero-shot and fine-tuned performance superiority on Emotional Support Conversation (ESC), compared with previous retrieval works. Our study suggests the potential to employ LLM as a foundation for a wider scope of retrieval tasks. Our codes, models, and datasets are available on https://github.com/flyfree5/LaHoRe.

Figures

Figures reproduced from arXiv: 2412.16615 by the authors.

**Figure 1.** Comparison of retrieval paradigms. Conventional fact-based retrieval tasks (Left) can be studied by query-document semantic similarity, while rationale-based retrieval tasks (Right) such as reply strategy or embodied subtask retrieval can not. However, wide applications of RAG systems sometimes challenge LLM for higher levels of cognitive capabilities, in which query and document are semantically related but not pa… view at source ↗

**Figure 2.** Architectural Comparison of LaHoRe with previous LLM-based retrieval methods. C, G, sim denote contrastive loss, generative loss, similarity function, respectively. (A): Bi-encoder; (B): Cross-encoder; (C): Generative RAG framework; (D): Framework of LaHoRe. We switch the order of Q and D such that D’s encoding can be cached. 2.2 Inference The relevance score is deduced from the positive-negative relative confidenc… view at source ↗

**Figure 3.** Retrieval performance plots with different positive-negative ratios. RQ2: With and without Ibc exhibit a huge difference which indicates its effectiveness. RQ3: Performance ablation on D⇒Q is relative balanced. However, response time (RT) can be improved apparently with D⇒Q where the prefix-cache mechanism applies. Sensitivity Analysis: In conventional embedding-based retrieval, it is common to construct in-batch … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 11 canonical work pages

[1]

In: arXiv preprint arXiv:2204.01691 (2022)

Ahn, M., Brohan, A., Brown, N., Chebotar, Y., Cortes, O., David, B., Finn, C., Fu, C., Gopalakrishnan, K., Hausman, K., Herzog, A., Ho, D., Hsu, J., Ibarz, J., Ichter, B., Irpan, A., Jang, E., Ruano, R.J., Jeffrey, K., Jesmonth, S., Joshi, N., Julian, R., Kalashnikov, D., Kuang, Y., Lee, K.H., Levine, S., Lu, Y., Luu, L., Parada, C., Pastor, P., Quiambao,...

arXiv 2022
[2]

In: Rogers, A., Boyd-Graber, J., Okazaki, N

Asai, A., Schick, T., Lewis, P., Chen, X., Izacard, G., Riedel, S., Hajishirzi, H., Yih, W.t.: Task-aware retrieval with instructions. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Findings of the Association for Computational Linguistics: ACL

work page
[3]

In: Goldberg, Y., Kozareva, Z., Zhang, Y

Cheng, Y., Liu, W., Li, W., Wang, J., Zhao, R., Liu, B., Liang, X., Zheng, Y.: Improving multi-turn emotional support dialogue generation with lookahead strat- egy planning. In: Goldberg, Y., Kozareva, Z., Zhang, Y. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. pp. 3014–

work page 2022
[4]

In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Fan, W., Ding, Y., Ning, L., Wang, S., Li, H., Yin, D., Chua, T.S., Li, Q.: A survey on rag meeting llms: Towards retrieval-augmented large language models. In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. p. 6491–6501. KDD ’24, Association for Computing Machin- ery, New York, NY, USA (2024). https://doi.org/10.114...

arXiv 2024
[5]

Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, M., Wang, H.: Retrieval-augmented generation for large language models: A survey (2024), https://arxiv.org/abs/2312.10997

arXiv 2024
[6]

In: arXiv preprint arXiv:2207.05608 (2022)

Huang, W., Xia, F., Xiao, T., Chan, H., Liang, J., Florence, P., Zeng, A., Tompson, J., Mordatch, I., Chebotar, Y., Sermanet, P., Brown, N., Jackson, T., Luu, L., Levine, S., Hausman, K., Ichter, B.: Inner monologue: Embodied reasoning through planning with language models. In: arXiv preprint arXiv:2207.05608 (2022)

arXiv 2022
[7]

In: Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles (2023)

Kwon, W., Li, Z., Zhuang, S., Sheng, Y., Zheng, L., Yu, C.H., Gonzalez, J.E., Zhang, H., Stoica, I.: Efficient memory management for large language model serv- ing with pagedattention. In: Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles (2023)

2023
[8]

In: Ku, L.W., Martins, A., Srikumar, V

Li, C., Liu, Z., Xiao, S., Shao, Y., Lian, D.: Llama2Vec: Unsupervised adaptation of large language models for dense retrieval. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Proceedings of the 62nd Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers). pp. 3490–3500. Association for Computational Linguistics, Bangkok, ...

work page 2024

Show all 34 references

[9]

org/abs/2404.14851 Large Language Model for Hidden Rationale-Based Retrieval 9

Li, X., Jin, J., Zhou, Y., Zhang, Y., Zhang, P., Zhu, Y., Dou, Z.: From matching to generation: A survey on generative information retrieval (2024),https://arxiv. org/abs/2404.14851 Large Language Model for Hidden Rationale-Based Retrieval 9

2024 arXiv
[10]

In: Zong, C., Xia, F., Li, W., Navigli, R

Liu, S., Zheng, C., Demasi, O., Sabour, S., Li, Y., Yu, Z., Jiang, Y., Huang, M.: Towards emotional support dialog systems. In: Zong, C., Xia, F., Li, W., Navigli, R. (eds.) Proceedings of the 59th Annual Meeting of the Association for Com- putational Linguistics and the 11th ...

2021 doi
[11]

In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

Ma, X., Wang, L., Yang, N., Wei, F., Lin, J.: Fine-tuning llama for multi-stage text retrieval. In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. p. 2421–2425. SIGIR ’24, Association for Computing Machinery, Ne...

2024
[12]

arXiv preprint arXiv:2202.08904 (2022)

Muennighoff, N.: Sgpt: Gpt sentence embeddings for semantic search. arXiv preprint arXiv:2202.08904 (2022)

2022 arXiv
[13]

Muennighoff, N., Su, H., Wang, L., Yang, N., Wei, F., Yu, T., Singh, A., Kiela, D.: Generative representational instruction tuning (2024)

2024
[14]

Nguyen, X.P., Pandit, S., Purushwalkam, S., Xu, A., Chen, H., Ming, Y., Ke, Z., Savarese, S., Xong, C., Joty, S.: Sfr-rag: Towards contextually faithful llms (2024), https://arxiv.org/abs/2409.09916

2024 arXiv
[15]

Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C.L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., Schulman, J., Hilton, J., Kelton, F., Miller, L., Simens, M., Askell, A., Welinder, P., Christiano, P., Leike, J., Lowe, R.: Training language models to follo...

2022 arXiv
[16]

Qiu, H., Li, A., Ma, L., Lan, Z.: Psychat: A client-centric dialogue system for mental health support (2024),https://arxiv.org/abs/2312.04262

2024 arXiv
[17]

Technical report, Alibaba Group (2024)

Qwen Team, A.G.: QWEN2 TECHNICAL REPORT. Technical report, Alibaba Group (2024)

2024
[18]

In: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S

Rafailov, R., Sharma, A., Mitchell, E., Manning, C.D., Ermon, S., Finn, C.: Di- rect preference optimization: Your language model is secretly a reward model. In: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S. (eds.) Ad- vances in Neural Information Proce...

2023
[19]

In: Proceedings of the Conference on Robot Learning (CoRL) (2023)

Ren, A.Z., Dixit, A., Bodrova, A., Singh, S., Tu, S., Brown, N., Xu, P., Takayama, L., Xia, F., Varley, J., Xu, Z., Sadigh, D., Zeng, A., Majumdar, A.: Robots that ask for help: Uncertainty alignment for large language model planners. In: Proceedings of the Conference on Robot...

2023
[20]

Robertson, S., Zaragoza, H.: The probabilistic relevance framework: Bm25 and beyond. Found. Trends Inf. Retr.3(4), 333–389 (Apr 2009). https://doi.org/ 10.1561/1500000019, https://doi.org/10.1561/1500000019

2009 doi
[21]

In: Rogers, A., Boyd-Graber, J., Okazaki, N

Su, H., Shi, W., Kasai, J., Wang, Y., Hu, Y., Ostendorf, M., Yih, W.t., Smith, N.A., Zettlemoyer, L., Yu, T.: One embedder, any task: Instruction- finetuned text embeddings. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Findings of the Association for Computational Lingu...

2023 doi
[22]

In: Zong, C., Xia, F., Li, W., Navigli, R

Sun, H., Lin, Z., Zheng, C., Liu, S., Huang, M.: PsyQA: A Chinese dataset for gen- erating long counseling text for mental health support. In: Zong, C., Xia, F., Li, W., Navigli, R. (eds.) Findings of the Association for Computational Linguistics: ACL- IJCNLP 2021. pp. 1489–15...

2021 doi
[23]

In: Ku, L.W., Martins, A., Srikumar, V

Wang, L., Yang, N., Huang, X., Yang, L., Majumder, R., Wei, F.: Improving text embeddings with large language models. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Proceedings of the 62nd Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers...

2024
[24]

Weller, O., Durme, B.V., Lawrie, D., Paranjape, A., Zhang, Y., Hessel, J.: Promptriever: Instruction-trained retrievers can be prompted like language models (2024), https://arxiv.org/abs/2409.11136

2024 arXiv
[25]

Yang, L., Yu, Z., Zhang, T., Cao, S., Xu, M., Zhang, W., Gonzalez, J.E., Cui, B.: Buffer of thoughts: Thought-augmented reasoning with large language models (2024), https://arxiv.org/abs/2406.04271

2024 arXiv
[26]

NAACL-HLT pp

Zhang, H., Diao, S., Lin, Y., Fung, Y.R., Lian, Q., Wang, X., Chen, Y., Ji, H., Zhang, T.: R-tuning: Instructing large language models to say ‘i don’t know’. NAACL-HLT pp. 7113–7139 (2024)

2024
[27]

Zhang, J., Peng, C., Sun, M., Chen, X., Liang, L., Zhang, Z., Zhou, J., Chen, H., Zhang, N.: Onegen: Efficient one-pass unified generation and retrieval for llms (2024), https://arxiv.org/abs/2409.05152

2024 arXiv
[28]

Zhao,P.,Zhang,H.,Yu,Q.,Wang,Z.,Geng,Y.,Fu,F.,Yang,L.,Zhang,W.,Jiang, J., Cui, B.: Retrieval-augmented generation for ai-generated content: A survey (2024), https://arxiv.org/abs/2402.19473

2024 arXiv
[29]

Zhao, S., Yang, Y., Wang, Z., He, Z., Qiu, L.K., Qiu, L.: Retrieval augmented generation (rag) and beyond: A comprehensive survey on how to make your llms use external data more wisely (2024),https://arxiv.org/abs/2409.14924

2024 arXiv
[30]

In: Rogers, A., Boyd-Graber, J., Okazaki, N

Zhao, W., Zhao, Y., Wang, S., Qin, B.: TransESC: Smoothing emotional sup- port conversation via turn-level state transition. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Findings of the Association for Computational Linguistics: ACL
[31]

arXiv preprint arXiv:2403.13372 (2024), http://arxiv.org/abs/2403.13372

Zheng, Y., Zhang, R., Zhang, J., Ye, Y., Luo, Z., Ma, Y.: Llamafactory: Unified efficient fine-tuning of 100+ language models. arXiv preprint arXiv:2403.13372 (2024), http://arxiv.org/abs/2403.13372

2024 arXiv
[33]

6725–6739

pp. 6725–6739. Association for Computational Linguistics, Toronto, Canada (Jul 2023). https://doi.org/10.18653/v1/2023.findings-acl.420, https:// aclanthology.org/2023.findings-acl.420

2023 doi
[2023]

3650–3675

pp. 3650–3675. Association for Computational Linguistics, Toronto, Canada (Jul 2023). https://doi.org/10.18653/v1/2023.findings-acl.225, https:// aclanthology.org/2023.findings-acl.225

2023 doi
[3026]

https://doi.org/10.18653/v1/2022.emnlp-main.195, https: //aclanthology.org/2022.emnlp-main.195

Association for Computational Linguistics, Abu Dhabi, United Arab Emi- rates (Dec 2022). https://doi.org/10.18653/v1/2022.emnlp-main.195, https: //aclanthology.org/2022.emnlp-main.195

2022 doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

In: arXiv preprint arXiv:2204.01691 (2022)

Ahn, M., Brohan, A., Brown, N., Chebotar, Y., Cortes, O., David, B., Finn, C., Fu, C., Gopalakrishnan, K., Hausman, K., Herzog, A., Ho, D., Hsu, J., Ibarz, J., Ichter, B., Irpan, A., Jang, E., Ruano, R.J., Jeffrey, K., Jesmonth, S., Joshi, N., Julian, R., Kalashnikov, D., Kuang, Y., Lee, K.H., Levine, S., Lu, Y., Luu, L., Parada, C., Pastor, P., Quiambao,...

arXiv 2022

[2] [2]

In: Rogers, A., Boyd-Graber, J., Okazaki, N

Asai, A., Schick, T., Lewis, P., Chen, X., Izacard, G., Riedel, S., Hajishirzi, H., Yih, W.t.: Task-aware retrieval with instructions. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Findings of the Association for Computational Linguistics: ACL

work page

[3] [3]

In: Goldberg, Y., Kozareva, Z., Zhang, Y

Cheng, Y., Liu, W., Li, W., Wang, J., Zhao, R., Liu, B., Liang, X., Zheng, Y.: Improving multi-turn emotional support dialogue generation with lookahead strat- egy planning. In: Goldberg, Y., Kozareva, Z., Zhang, Y. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. pp. 3014–

work page 2022

[4] [4]

In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Fan, W., Ding, Y., Ning, L., Wang, S., Li, H., Yin, D., Chua, T.S., Li, Q.: A survey on rag meeting llms: Towards retrieval-augmented large language models. In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. p. 6491–6501. KDD ’24, Association for Computing Machin- ery, New York, NY, USA (2024). https://doi.org/10.114...

arXiv 2024

[5] [5]

Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, M., Wang, H.: Retrieval-augmented generation for large language models: A survey (2024), https://arxiv.org/abs/2312.10997

arXiv 2024

[6] [6]

In: arXiv preprint arXiv:2207.05608 (2022)

Huang, W., Xia, F., Xiao, T., Chan, H., Liang, J., Florence, P., Zeng, A., Tompson, J., Mordatch, I., Chebotar, Y., Sermanet, P., Brown, N., Jackson, T., Luu, L., Levine, S., Hausman, K., Ichter, B.: Inner monologue: Embodied reasoning through planning with language models. In: arXiv preprint arXiv:2207.05608 (2022)

arXiv 2022

[7] [7]

In: Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles (2023)

Kwon, W., Li, Z., Zhuang, S., Sheng, Y., Zheng, L., Yu, C.H., Gonzalez, J.E., Zhang, H., Stoica, I.: Efficient memory management for large language model serv- ing with pagedattention. In: Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles (2023)

2023

[8] [8]

In: Ku, L.W., Martins, A., Srikumar, V

Li, C., Liu, Z., Xiao, S., Shao, Y., Lian, D.: Llama2Vec: Unsupervised adaptation of large language models for dense retrieval. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Proceedings of the 62nd Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers). pp. 3490–3500. Association for Computational Linguistics, Bangkok, ...

work page 2024