Pith. sign in

REVIEW 3 major objections 5 minor 37 references

TAPO: Task-Referenced Adaptation for Prompt Optimization

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read TAPO reports that letting each task choose its own evaluation metrics beats CoT, APE, PE2, and PromptBreeder on six datasets.

desk verdict TAPO has a genuinely new metric-selection idea, but its own table contradicts the 'consistently outperforms' claim and the reported metric is the same cosine similarity used in optimization, so the empirical case needs major revision. read the letter →

arxiv 2501.06689 v3 pith:N5GVS7EP submitted 2025-01-12 cs.CL

Wenxin Luo , Weirui Wang , Xiaopeng Li , Weibo Zhou , Pengyue Jia , Xiangyu Zhao This is my paper · ORCID
classification cs.CL
keywords automatedpromptoptimizationtask-awaremetricselectionevolutionarysearchmulti-metricevaluationlargelanguagemodelscosinesimilarityscoringchain-of-thoughtpromptingmathwordproblemreasoning
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

TAPO is an automated prompt optimization framework built around one idea: the right evaluation metric for a prompt depends on the task, so the optimizing LLM should choose and weight the metrics itself. For each dataset, the framework first classifies the task type, then selects and weights metrics from a pool of similarity, diversity, perplexity, and complexity, producing a task-specific scoring function. Candidate prompts are scored with that function, and the best ones are repeatedly mutated and re-selected through a tournament mechanism. On the paper's reported numbers, this scheme outperforms or matches CoT, APE, PE2, and PromptBreeder across BBH, GSM8K, AddSub, MultiArith, SingleEQ, and SVAMP on GPT-3.5-turbo and GPT-4o, and it improves Llama3-8B-Instruct's math reasoning scores by about 10.2% over CoT. The authors take the ablation results as evidence that both the multi-metric scoring and the prompt-optimization loop contribute to the gains.

What carries the argument

The load-bearing mechanism is the pairing of Dynamic Metric Selection with the weighted scoring function $S(P)=\sum_{i=1}^{n} w_i M_i(P)$, where the $w_i$ are assigned by the LLM according to task priority and the $M_i$ are similarity, diversity, perplexity, and complexity scores. That function turns 'adapt to the task' into a concrete objective that drives the third module, Evolution-Based Prompt Optimization, which applies small strategy mutations such as 'break the task into steps' and uses tournament selection to keep the highest-scoring prompts. The named machinery is the metric-selection-then-evolve loop, and the claim is that it transfers across math, reasoning, and language tasks.

What would settle it

Run TAPO with a strict split: optimize prompts on a training fold, validate on a validation fold, then evaluate on a never-seen test fold using exact-match or human-judged correctness instead of cosine similarity; if higher similarity scores do not accompany higher correctness on the held-out test fold, the central claim fails.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that prompt optimization benefits from making the fitness function task-dependent instead of fixed. TAPO operationalizes this with a three-module loop: an LLM-driven module selects task-relevant metrics and assigns weights, a scoring module combines them into $S(P)=\sum_{i=1}^{n} w_i M_i(P)$, and an evolution module mutates and tournament-selects the best prompts across iterations. The empirical core is Table I, where TAPO is either the top or second-best method in nearly every dataset-model pairing, with particularly large margins on arithmetic tasks such as AddSub and MultiArith under GPT-3.5-turbo. The paper interprets this as showing that dynamic metric selection plus evolutionary refinement yields prompts that are better adapted to each task's demands.

Load-bearing premise

The load-bearing premise is that the LLM-chosen metrics, especially MiniLM cosine similarity, and their weights in the scoring function faithfully measure task quality, so that optimizing prompts against them transfers to genuinely better task solving.

Editorial extensions

If this is right

  • The reported gains are largest on arithmetic reasoning: with GPT-3.5-turbo, TAPO reaches 88.15% on AddSub and 89.26% on MultiArith, versus 58.61% and 69.00% for zero-shot CoT.
  • The improvement is not universal: on GPT-4o GSM8K, TAPO scores 88.40% versus PromptBreeder's 88.61%, and on GPT-3.5-turbo SVAMP it scores 92.72% versus CoT's 94.38%, so the paper's claim is consistency near the top rather than outright dominance.
  • Ablations indicate both components matter: removing multi-metric scoring drops SingleEQ from 89.06% to 75.91%, and removing the prompt-optimization loop drops MultiArith from 89.26% to 83.04%.
  • On Llama3-8B-Instruct, TAPO improves average math-reasoning similarity scores by 10.2% over CoT and 6.2% over PE2, suggesting the method also helps weaker open models keep output format during multi-step reasoning.

Reading between the lines

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

  • Because MiniLM cosine similarity is both the optimization objective and the reported evaluation metric, part of the gain could be improved embedding resemblance rather than improved reasoning; a held-out exact-match accuracy check would separate the two.
  • The same weighted-metric loop could be applied to tasks without reference answers by replacing cosine similarity with self-consistency or entailment-based proxies, but TAPO as presented does not test those settings.
  • Since no train/validation/test split is described, the reported numbers may be optimistic; rerunning the pipeline with separate optimization and evaluation partitions is a direct way to measure that optimism.
  • Letting the LLM choose its own evaluation criteria introduces a possible gaming channel: the model may select metrics its own outputs happen to satisfy. A control condition with fixed or randomly chosen metrics would isolate how much of the gain comes from adaptation rather than from the evolutionary search itself.
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

3 major / 5 minor

Summary. This paper presents TAPO, a framework for automated prompt optimization in which an LLM classifies the task, selects evaluation metrics (similarity, diversity, perplexity, complexity) with weights, and then uses an evolutionary loop of mutation and tournament selection to refine prompts under the weighted scoring function S(P)=sum_i w_i*M_i(P). The authors compare TAPO against Zero-Shot CoT, APE, PE2, and PromptBreeder on BBH, GSM8K, AddSub, MultiArith, SingleEQ, and SVAMP using GPT-3.5-turbo, GPT-4o, and Llama3-8B-Instruct, and they report ablations that remove prompt optimization and multi-metric scoring. The main claim is that TAPO consistently outperforms baseline methods by dynamically selecting and weighting task-specific metrics.

Significance. If the empirical claim were supported, TAPO would be a useful contribution: it targets multi-task adaptation in prompt optimization, ships public code, includes ablations, and covers both closed and open models. However, the headline results are similarity scores computed with the same MiniLM cosine-similarity model that appears inside the optimization objective, and the paper gives no data split or uncertainty quantification. The contribution is therefore conditional on a re-evaluation with an objective-independent, held-out metric.

major comments (3)
  1. [§II-C, §III-A, Table I, Fig. 2] The evaluation metric in Table I and Figure 2 is described as cosine similarity between generated and reference texts using all-MiniLM-L6-v2, while Eq. (1) makes the same embedding-based similarity a component of S(P), the score used for tournament selection and mutation in §II-D. The reported numbers therefore measure the objective being optimized, so they cannot by themselves show that TAPO improves task-solving quality (exact answers, correct reasoning, etc.) rather than similarity to reference strings. Please report task-native metrics (e.g., exact-match accuracy) and/or a metric excluded from the optimization objective.
  2. [§III-A] The experiment settings do not describe any train/validation/test split. Without a held-out test set, prompt selection during evolution may be performed on the same examples used to compute the final reported scores, which would make the comparison reflect selection on the test set. Please specify how many examples are used for metric selection and prompt evolution, and evaluate the final prompts on a disjoint held-out set.
  3. [§III-B, Table I] The claim that TAPO consistently outperforms baseline methods is not supported by the point estimates in Table I: on GSM8K with GPT-4o, PromptBreeder scores 88.61 while TAPO scores 88.40, and on SVAMP with GPT-3.5-turbo, CoT scores 94.38 while TAPO scores 92.72. In addition, the asterisks for p<0.05 are not accompanied by variance, number of independent runs, or a description of the significance test. Please quantify uncertainty and either provide a precise definition of consistent or soften the claim.
minor comments (5)
  1. [Table III] The table header contains the typo 'SigleEQ'; it should be 'SingleEQ'.
  2. [Section I] The string 'V oke-k' appears to be a rendering artifact, and the correspondence to reference [4] is unclear; please correct the citation and the model name.
  3. [Section III-A, Table II] Table II includes a translation error detection example, but no translation dataset is listed in the datasets paragraph; please specify which of the six datasets or BBH tasks this example comes from.
  4. [Section II-A] The sentence 'Dynamically adapts to various tasks by selecting appropriate evaluation metrics...' lacks a grammatical subject; consider beginning 'TAPO dynamically adapts...'.
  5. [Figure 2] The figure caption and surrounding text do not state exactly which quantity is plotted (similarity score versus accuracy) or which datasets are aggregated in the reported average improvements; please clarify the axes and the aggregation procedure.

Circularity Check

1 steps flagged · score 6.0 of 10

TAPO's reported 'similarity scores' are a component of its own optimization objective, so the central outperformance claim partly reduces to optimizing the evaluation metric.

  1. fitted input called prediction [Section III-B (claim), Section II-C Eq. (1), Section III-A, Table I caption]
    "TAPO consistently outperforms baseline methods by dynamically selecting and weighting task-specific metrics. ... The scoring function is defined as: S(P) = nX i=1 wi · Mi(P) (1) ... Similarity is assessed using cosine similarity, calculated via the all-MiniLM-L6-v2 model, to measure the semantic alignment between the generated and reference texts. ... PERFORMANCE COMPARISON ON DIFFERENT DATASETS WITH GPT-3.5-TURBO AND GPT-4O (SIMILARITY SCORES)."

    TAPO evolves prompts by maximizing S(P), whose metric set includes cosine similarity against reference texts. The paper's headline results (Table I, Section III-B) are 'similarity scores' computed with that same all-MiniLM-L6-v2 cosine similarity. Thus the reported quantity is not an independent outcome measure: it is one of the terms directly optimized during prompt evolution. The 'prediction' that TAPO outperforms baselines is partly forced by the fitness function, and no separate task-accuracy or held-out evaluation is reported to break the loop.

full rationale

The only load-bearing circular step is the metric loop: the evaluation metric used to declare TAPO's success (cosine similarity / 'similarity score') is also a component of the objective S(P) in Eq. (1) that guides prompt selection and mutation. This makes the empirical claim in Section III-B partially self-referential. The paper does not describe a train/validation/test split, so selection may occur on the same examples used for reporting, compounding the issue. No load-bearing self-citation chain was found: the many same-lab references in the Related Work section are contextual and do not ground TAPO's core claim. The fact that Table I itself shows several baseline wins (e.g., CoT on GSM8K GPT-3.5-turbo and on AddSub GPT-4o) is a correctness/statistical concern rather than circularity. Overall, because the framework does contain an independent evolutionary component and the optimized objective is a weighted multi-metric score rather than the similarity score alone, the circularity is partial rather than total.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central empirical claim rests on several unproven modeling choices: the LLM's ability to pick valid metrics, the validity of the weighted metric sum, the validity of cosine similarity and perplexity as proxies for math task correctness, and the convergence of evolutionary search. These are domain assumptions, not standard mathematical axioms. The metric weights selected by the LLM function as unreported free parameters.

free parameters (1)
  • Metric weights w_i in Eq (1) = Not reported; assigned by the LLM per task in Dynamic Metric Selection
    The scoring function S(P) = sum w_i M_i(P) depends on LLM-assigned weights that are never disclosed or ablated for sensitivity; they determine which prompts survive evolution and thus affect every reported number.
assumptions (4)
  • domain assumption The LLM can reliably identify task type and select relevant evaluation metrics for that task (Section II-B).
    Dynamic Metric Selection provides no evaluation of whether the selected metrics correlate with actual task performance; the whole method assumes this correlation.
  • domain assumption The weighted sum in Eq. (1) is a valid objective for prompt quality (Section II-C).
    No evidence is given that a linear combination of similarity, diversity, perplexity, and complexity correlates with human-judged or downstream task quality.
  • domain assumption Cosine similarity from all-MiniLM-L6-v2 and GPT-2 perplexity are appropriate proxies for correctness and fluency in arithmetic and reasoning tasks (Section III-A).
    Math word problems are judged by numeric correctness, not semantic similarity; using similarity as the reported performance metric may not measure the intended capability.
  • domain assumption The mutation strategy library and tournament selection converge to better prompts within the iteration limit (Section II-D).
    No convergence analysis or sensitivity study is provided; the improvement is demonstrated only empirically on selected datasets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TAPO: Task-Referenced Adaptation for Prompt Optimization." pith.science (2026). https://pith.science/paper/N5GVS7EP

@misc{pith2026250106689,
  author       = {Pith},
  title        = {Pith review of: TAPO: Task-Referenced Adaptation for Prompt Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N5GVS7EP}},
  note         = {Machine review of arXiv:2501.06689}
}
read the original abstract

Prompt engineering can significantly improve the performance of large language models (LLMs), with automated prompt optimization (APO) gaining significant attention due to the time-consuming and laborious nature of manual prompt design. However, much of the existing work in APO overlooks task-specific characteristics, resulting in prompts that lack domain specificity and are not well-suited for task-specific optimization. In this paper, we introduce TAPO, a multitask-aware prompt optimization framework composed of three key modules. First, a task-aware metric selection module is proposed to enhance task-specific prompt generation capabilities. Second, we present a multi-metrics evaluation module to jointly evaluate prompts from multiple perspectives. Third, an evolution-based optimization framework is introduced for automatic prompt refinement, which improves adaptability across various tasks. Extensive experiments on six datasets demonstrate the effectiveness of our approach, and our code is publicly available.

Figures

Figures reproduced from arXiv: 2501.06689 by the authors.

Figure 1
Figure 1. The framework of TAPO. For Dynamic Metric Selection, We provide a task dataset example for the LLM to select metrics and assign weights based on priority, creating task-specific evaluation metrics for Task-Aware Prompt Evaluation. We employ a tournament selection algorithm for Evolution-Based Prompt Optimization to select and mutate the better-performing prompts, adding task-adapted prompts to the candidates. refini… view at source ↗
Figure 2
Figure 2. Performance Comparison with Llama3-8B-Instruct. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 10 canonical work pages

  1. [1]

    A survey of large language models,

    W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Dong et al. , “A survey of large language models,” arXiv preprint arXiv:2303.18223, 2023

    arXiv 2023

  2. [2]

    Large language models are human-level prompt engineers,

    Y . Zhou, A. I. Muresanu, Z. Han, K. Paster, S. Pitis, H. Chan, and J. Ba, “Large language models are human-level prompt engineers,”arXiv preprint arXiv:2211.01910, 2022

    arXiv 2022

  3. [3]

    Tem- pera: Test-time prompting via reinforcement learning,

    T. Zhang, X. Wang, D. Zhou, D. Schuurmans, and J. E. Gonzalez, “Tem- pera: Test-time prompting via reinforcement learning,” arXiv preprint arXiv:2211.11890, 2022

    arXiv 2022

  4. [4]

    Selective annota- tion makes language models better few-shot learners,

    H. Su, J. Kasai, C. H. Wu, W. Shi, T. Wang, J. Xin, R. Zhang, M. Ostendorf, L. Zettlemoyer, N. A. Smith et al. , “Selective annota- tion makes language models better few-shot learners,” arXiv preprint arXiv:2209.01975, 2022

    arXiv 2022

  5. [5]

    Automatic chain of thought prompting in large language models,

    Z. Zhang, A. Zhang, M. Li, and A. Smola, “Automatic chain of thought prompting in large language models,” arXiv preprint arXiv:2210.03493, 2022

    arXiv 2022

  6. [6]

    Promptbreeder: Self-referential self-improvement via prompt evolution,

    C. Fernando, D. Banarse, H. Michalewski, S. Osindero, and T. Rockt ¨aschel, “Promptbreeder: Self-referential self-improvement via prompt evolution,” arXiv preprint arXiv:2309.16797 , 2023

    arXiv 2023

  7. [7]

    Eliciting human preferences with language models,

    B. Z. Li, A. Tamkin, N. Goodman, and J. Andreas, “Eliciting human preferences with language models,” arXiv preprint arXiv:2310.11589 , 2023

    arXiv 2023

  8. [8]

    Machine Translation with Large Language Models: Prompt Engineering for Persian, English, and Russian Directions

    S. E. S. Nooshin Pourkamali, “Machine translation with large language models: Prompt engineering for persian, english, and russian directions,” arXiv preprint arXiv:2401.08429 , 2024

    work page Pith review arXiv 2024

Show all 37 references
  1. [9]

    Learning to solve arithmetic word problems with verb categorization,

    M. J. Hosseini, H. Hajishirzi, O. Etzioni, and N. Kushman, “Learning to solve arithmetic word problems with verb categorization,” in Proceed- ings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014, pp. 523–533

    2014

  2. [10]

    Solving general arithmetic word problems,

    S. Roy and D. Roth, “Solving general arithmetic word problems,” arXiv preprint arXiv:1608.01413, 2016

    2016 arXiv

  3. [11]

    Parsing algebraic word problems into equations,

    R. Koncel-Kedziorski, H. Hajishirzi, A. Sabharwal, O. Etzioni, and S. D. Ang, “Parsing algebraic word problems into equations,” Transactions of the Association for Computational Linguistics , vol. 3, pp. 585–597, 2015

    2015

  4. [12]

    Are nlp models really able to solve simple math word problems?

    A. Patel, S. Bhattamishra, and N. Goyal, “Are nlp models really able to solve simple math word problems?” arXiv preprint arXiv:2103.07191 , 2021

    2021 arXiv

  5. [13]

    Training verifiers to solve math word problems,

    K. Cobbe, V . Kosaraju, M. Bavarian, M. Chen, H. Jun, L. Kaiser, M. Plappert, J. Tworek, J. Hilton, R. Nakano et al., “Training verifiers to solve math word problems,” arXiv preprint arXiv:2110.14168 , 2021

    2021 arXiv

  6. [14]

    Challenging big-bench tasks and whether chain-of-thought can solve them,

    M. Suzgun, N. Scales, S. Gehrmann et al. , “Challenging big-bench tasks and whether chain-of-thought can solve them,” arXiv preprint arXiv:2210.09261, 2022

    2022 arXiv

  7. [15]

    Chain-of-thought prompting elicits reasoning in large language models,

    J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. Chi, Q. V . Le, D. Zhou et al. , “Chain-of-thought prompting elicits reasoning in large language models,” Advances in neural information processing systems , vol. 35, pp. 24 824–24 837, 2022

    2022

  8. [16]

    Prompt engineering a prompt engineer,

    Q. Ye, M. Axmed, R. Pryzant, and F. Khani, “Prompt engineering a prompt engineer,” arXiv preprint arXiv:2311.05661 , 2023

    2023 arXiv

  9. [17]

    Gpt-3.5 turbo model documentation,

    OpenAI, “Gpt-3.5 turbo model documentation,” https://platform.openai. com/docs/models/gpt-3-5#gpt-3-5-turbo, 2023

    2023

  10. [18]

    Gpt-4o system card,

    A. Hurst, A. Lerer, A. P. Goucher, A. Perelman, A. Ramesh, A. Clark, A. Ostrow, A. Welihinda, A. Hayes, A. Radford et al., “Gpt-4o system card,” arXiv preprint arXiv:2410.21276 , 2024

    2024 arXiv

  11. [19]

    The llama 3 herd of models,

    A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan et al. , “The llama 3 herd of models,” arXiv preprint arXiv:2407.21783 , 2024

    2024 arXiv

  12. [20]

    Sentence-bert: Sentence embeddings using siamese bert- networks,

    N. Reimers, “Sentence-bert: Sentence embeddings using siamese bert- networks,” arXiv preprint arXiv:1908.10084 , 2019

    1908 arXiv

  13. [21]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

    2019

  14. [22]

    Agentir: 1st workshop on agent-based information retrieval,

    Q. Cai, X. Zhao, L. Pan, X. Xin, J. Huang, W. Zhang, L. Zhao, D. Yin, and G. H. Yang, “Agentir: 1st workshop on agent-based information retrieval,” in Proceedings of the 47th International ACM SIGIR Confer- ence on Research and Development in Information Retrieval , 2024, pp. ...

    2024

  15. [23]

    Agent4ranking: Semantic robust ranking via personalized query rewrit- ing using multi-agent llm,

    X. Li, L. Su, P. Jia, X. Zhao, S. Cheng, J. Wang, and D. Yin, “Agent4ranking: Semantic robust ranking via personalized query rewrit- ing using multi-agent llm,” arXiv preprint arXiv:2312.15450 , 2023

    2023 arXiv

  16. [24]

    Bridging relevance and reasoning: Rationale distillation in retrieval-augmented generation,

    P. Jia, D. Xu, X. Li, Z. Du, X. Li, X. Zhao, Y . Wang, Y . Wang, H. Guo, and R. Tang, “Bridging relevance and reasoning: Rationale distillation in retrieval-augmented generation,” arXiv preprint arXiv:2412.08519 , 2024

    2024

  17. [25]

    G3: an effective and adaptive framework for worldwide geolocalization using large multi-modality models,

    P. Jia, Y . Liu, X. Li, Y . Wang, Y . Du, X. Han, X. Wei, S. Wang, D. Yin, and X. Zhao, “G3: an effective and adaptive framework for worldwide geolocalization using large multi-modality models,” arXiv preprint arXiv:2405.14702, 2024

    2024 arXiv

  18. [26]

    Syneg: Llm-driven synthetic hard-negatives for dense retrieval,

    X. Li, X. Li, H. Zhang, Z. Du, P. Jia, Y . Wang, X. Zhao, H. Guo, and R. Tang, “Syneg: Llm-driven synthetic hard-negatives for dense retrieval,” arXiv preprint arXiv:2412.17250 , 2024

    2024 arXiv

  19. [27]

    Mill: Mutual verification with large language models for zero-shot query expansion,

    P. Jia, Y . Liu, X. Zhao, X. Li, C. Hao, S. Wang, and D. Yin, “Mill: Mutual verification with large language models for zero-shot query expansion,” arXiv preprint arXiv:2310.19056 , 2023

    2023 arXiv

  20. [28]

    When moe meets llms: Parameter efficient fine-tuning for multi-task medical applications,

    Q. Liu, X. Wu, X. Zhao, Y . Zhu, D. Xu, F. Tian, and Y . Zheng, “When moe meets llms: Parameter efficient fine-tuning for multi-task medical applications,” in Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval , 2024...

    2024

  21. [29]

    Llm-esr: Large language models enhancement for long-tailed sequential recommendation,

    Q. Liu, X. Wu, Y . Wang, Z. Zhang, F. Tian, Y . Zheng, and X. Zhao, “Llm-esr: Large language models enhancement for long-tailed sequential recommendation,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems , 2024

    2024

  22. [30]

    Hamur: Hyper adapter for multi-domain recommendation,

    X. Li, F. Yan, X. Zhao, Y . Wang, B. Chen, H. Guo, and R. Tang, “Hamur: Hyper adapter for multi-domain recommendation,” in Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 2023, pp. 1268–1277

    2023

  23. [31]

    Hierrec: Scenario-aware hierarchical modeling for multi-scenario recommendations,

    J. Gao, B. Chen, M. Zhu, X. Zhao, X. Li, Y . Wang, Y . Wang, H. Guo, and R. Tang, “Hierrec: Scenario-aware hierarchical modeling for multi-scenario recommendations,” in Proceedings of the 33rd ACM International Conference on Information and Knowledge Management , 2024, pp. 653–662

    2024

  24. [32]

    D3: A methodological exploration of domain division, modeling, and balance in multi-domain recommendations,

    P. Jia, Y . Wang, S. Lin, X. Li, X. Zhao, H. Guo, and R. Tang, “D3: A methodological exploration of domain division, modeling, and balance in multi-domain recommendations,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 8, 2024, pp. 8553– 8561

    2024

  25. [33]

    Scenario-wise rec: A multi-scenario recommendation benchmark,

    X. Li, J. Gao, P. Jia, Y . Wang, W. Wang, Y . Wang, Y . Wang, H. Guo, and R. Tang, “Scenario-wise rec: A multi-scenario recommendation benchmark,” arXiv preprint arXiv:2412.17374 , 2024

    2024

  26. [34]

    Large language model enhanced recommender systems: Taxonomy, trend, application and future,

    Q. Liu, X. Zhao, Y . Wang, Y . Wang, Z. Zhang, Y . Sun, X. Li, M. Wang, P. Jia, C. Chen et al. , “Large language model enhanced recommender systems: Taxonomy, trend, application and future,” arXiv preprint arXiv:2412.13432, 2024

    2024 arXiv

  27. [35]

    Llm-powered user simulator for recommender system,

    Z. Zhang, S. Liu, Z. Liu, R. Zhong, Q. Cai, X. Zhao, C. Zhang, Q. Liu, and P. Jiang, “Llm-powered user simulator for recommender system,” arXiv preprint arXiv:2412.16984 , 2024

    2024 arXiv

  28. [36]

    Rlprompt: Optimizing discrete text prompts with reinforcement learning,

    M. Deng, J. Wang, and Hsieh, “Rlprompt: Optimizing discrete text prompts with reinforcement learning,” arXiv preprint arXiv:2205.12548, 2022

    2022 arXiv

  29. [37]

    Large language models to enhance bayesian optimization,

    T. Liu, N. Astorga, N. Seedat, and M. van der Schaar, “Large language models to enhance bayesian optimization,” arXiv preprint arXiv:2402.03921, 2024

    2024 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.