Pith. sign in

REVIEW 4 major objections 6 minor 50 references

Learning by Analogy: Enhancing Few-Shot Prompting for Math Word Problem Solving with Computational Graph-Based Retrieval

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A retriever that matches the arithmetic structure of math word problems improves few-shot exact-match accuracy by up to 6.7 points on average over semantic baselines, and up to 19.5 points over random selection, without retraining the…

desk verdict Computational-graph-based exemplar retrieval is a real, consistent win for few-shot math word problems, but the paper's mechanistic claim that graph analogy drives the gains is only directly validated on the training distribution. read the letter →

arxiv 2411.16454 v1 pith:KTOHJ3AZ submitted 2024-11-25 cs.CL

Xiaocong Yang , Jiacheng Lin , Ziqi Wang , Chengxiang Zhai This is my paper · ORCID
classification cs.CL
keywords mathwordproblemsfew-shotpromptingcomputationalgraphretrieval-augmentedgenerationcontrastivelearninglargelanguagemodelsexact-matchaccuracyanalogy-basedretrieval
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

This paper argues that the bottleneck in few-shot math word problem solving is not the generator but the choice of examples: randomly picked or semantically similar exemplars often share surface wording while differing in the arithmetic operations needed. The authors train a lightweight text retriever with contrastive learning, using questions that share the same computational graph as positive pairs, and then feed the top-k retrieved question-answer pairs into the prompt. Across six English and Chinese datasets and five generator sizes, this structural retrieval outperforms semantic retrieval by up to 6.7 points on average exact-match accuracy, and random selection by up to 19.5 points, without fine-tuning the generator. The result matters because it offers a modular, training-free way to strengthen LLM reasoning on math: give the model an analogous worked solution rather than a merely topical one.

What carries the argument

The carrying object is the computational graph: a normalized arithmetic expression (e.g., $x=\text{num}_a*\text{num}_b/\text{num}_c$) that abstracts away numbers, entities, and wording to capture the reasoning path of a math word problem. The mechanism around it is a contrastively trained retriever: an encoder with an added two-layer MLP pooler, trained with an infoNCE loss where positive pairs are questions with identical computational graphs and negatives are in-batch questions with different graphs. At inference the retriever computes top-$k$ similar questions from the corpus and their solutions are concatenated into the few-shot prompt. This machinery does the work of aligning example selection with the hidden structure of the target problem, without retraining or altering the generation model.

What would settle it

A decisive check: on a graph-labeled test set, compare prompts built from top-k retrievals whose graphs match the query against prompts built from top-k retrievals whose graphs do not match but whose texts are equally similar; if exact-match accuracy is no higher in the graph-matched condition, the paper's central mechanism is not the cause of the gains.

Watch

Extended reading notes

Core claim

The central claim is that analogy by arithmetic structure, not by topic, is what makes few-shot exemplars useful for math word problems. The paper represents each problem's reasoning path as a computational graph, a template such as $x=(\text{num}_a - \text{num}_b)/\text{num}_c$, and trains an encoder-based retriever so that questions whose graphs are identical embed close together. At inference, the retriever selects the top-$k$ question-solution pairs from the corpus by embedding similarity and places them in the prompt; the generator is untouched. On Math23k, ape210k, gsm8k, math_qa, Calc-ape210k, and aqua_rat, this recipe improves exact-match accuracy over semantic retrieval across nearly all settings, with the largest gains on smaller generators, and approaches the upper bound obtained when the true graph itself is used for retrieval on Math23k.

Load-bearing premise

The load-bearing premise is that a retriever trained on exact computational-graph equality within 25% of Math23k transfers structural similarity to five other datasets, so that top-$k$ text embeddings select exemplars whose hidden arithmetic structure matches the query rather than merely their wording.

Editorial extensions

If this is right

  • On five of the six datasets, choosing exemplars by computational-graph analogy improves exact-match accuracy over both semantic and random retrieval across generator sizes from 0.5B to 70B/72B.
  • The gain is largest for small generators, so structural retrieval can partially compensate for weaker reasoning in compact models.
  • Training on as little as 25% of Math23k (about 5,000 samples) is enough to reach stable retrieval quality, and 10% already yields a 4-point gain over semantic retrieval on aqua_rat.
  • The retriever still helps when the retrieval corpus is out-of-distribution: using ape210k as the corpus for SuperCLUE-Math6 test questions beats semantic and random selection.
  • LLM-distilled positive pairs generated by rewriting questions to keep the same graph can substitute for human-labeled graphs, outperforming semantic retrieval while slightly underperforming labeled training data.

Reading between the lines

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

  • Beyond the six datasets tested, the same graph-analogy retrieval could plausibly generalize to other structured reasoning tasks—such as symbolic logic, table-based arithmetic, or code—where a problem's 'reasoning path' can be normalized into an abstract template.
  • The paper's own observation that distilled data keeps sentence structure more similar than labeled data suggests an implicit risk: a text retriever trained on such pairs may latch onto syntactic templates rather than true graphs; one testable fix is to adversarially vary sentence structure more aggressively during distillation.
  • Because gains concentrate in smaller models, an implicit consequence is that structural retrieval is a cheap way to deploy capable math solving on edge devices, where large generators are impractical.
  • The upper-bound experiment implies the ceiling is not the generator but the retriever's ability to approximate graph similarity; further gains might come from combining graph-level supervision with more powerful encoders, not from larger prompts.
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 a retrieval-augmented few-shot prompting method for math word problems (MWPs). The retriever is an encoder trained with a contrastive objective on Math23k question pairs whose computational graph templates are identical; at inference it retrieves the top-k most similar question-solution pairs from the training corpus and prepends them to the query as few-shot exemplars for an LLM generator. The generator is not retrained. The authors evaluate on six MWP datasets (Math23k, ape210k, gsm8k, math_qa, Calc-ape210k, aqua_rat) with generators from 0.5B to 70B parameters, reporting exact-match accuracy. They report average gains of up to 6.7 points over BGE retrieval and 19.5 points over random selection, plus analyses of an upper bound on Math23k, a qualitative case study, a correlation scatter plot, training-data scaling, corpus mismatch, and LLM-distilled training data without human labels.

Significance. If the empirical gains are reliable, the method is practically attractive: it is modular, requires no generator fine-tuning, and the analysis in Section 3.3.3 suggests that only a few thousand labeled pairs suffice. The paper also makes a good-faith effort to study the mechanism, including a direct upper-bound comparison on Math23k and a corpus-shift experiment, and it is transparent about the limitations of the distilled-data variant. The main weakness is evidentiary: all accuracy numbers are single runs, and the central claim that the retriever selects structurally analogous exemplars is directly validated only on the retriever's own training distribution. The current evidence supports the conclusion that fine-tuning the retriever on math text improves few-shot prompting, but it does not yet establish that computational-graph structure, rather than improved semantic/domain matching, is the active ingredient for the out-of-distribution datasets.

major comments (4)
  1. [Section 3.2, Table 2] All results in Table 2 are single runs; no seeds, standard deviations, confidence intervals, or significance tests are reported. Several per-dataset differences are small or negative: gsm8k with LLaMA-3B is +0.4 over BGE, and math_qa is -0.7, -0.6, and -0.5 for the 0.5B, 3B, and 70B generators, respectively. Without variance estimates, the abstract's claim of 'significant improvement' and the headline 6.7-point average cannot be properly assessed; the average also hides multiple datasets where the proposed method does not improve over BGE.
  2. [Section 3.3.2, Figure 4, and Table 2] The central mechanism claim is that the trained retriever selects exemplars whose computational graphs match the query's graph. This is directly quantified only on Math23k (Figure 4), which is the retriever's training distribution. For ape210k, gsm8k, math_qa, Calc-ape210k, and aqua_rat, no quantitative measure of graph similarity between queries and retrieved exemplars is reported; the Calc-ape210k case study (Figure 3) is qualitative. Since the retriever is fine-tuned on math text, the observed gains in Table 2 could arise from improved semantic or domain-specific retrieval rather than from computational-graph analogy. This is load-bearing because the claimed advantage over BGE rests on structural alignment. Please report graph-template overlap or normalized graph-string similarity between query and retrieved exemplars on datasets with equation solutions (ape210k, Calc-ape210k), or include a control retriever fine-tuned on math-domain contrastive pairs without graph labels to show that the gains require the graph-based positive pairs.
  3. [Section 4] The authors state that in the distilled-data training, 'the sentence structure before and after rewriting is more similar than in the labeled data pairs, which the retriever may rely on to capture similarity between positive pairs during training, rather than focusing on the true computational graphs.' This is a self-admitted limitation of the Section 4 pipeline, and it directly undermines the claim that the distilled-data retriever works by graph analogy. It also reinforces the concern in the previous comment: the retriever can improve downstream accuracy without necessarily encoding graph structure. The Section 4 claim of 'computational graph-free' training data should be qualified, and the sentence-structure confound should be measured, for example by comparing lexical or syntactic overlap between positive pairs in the labeled and distilled settings.
  4. [Section 3.3.1, Table 3] The label 'upper bound' for retrieval by normalized Levenshtein distance on graph strings is not strictly justified. It assumes that graph-similar retrieval is the optimal retrieval policy for the generator, but the generator could in principle benefit more from some other selection criterion. The comparison is still informative as a sanity check on retriever quality, but the terminology overclaims and should be softened, for example to 'graph-oracle reference' or 'graph-retrieval reference'.
minor comments (6)
  1. [Table 2] The column header 'A vg.' appears to be a typo for 'Avg.'.
  2. [Figure 2 and Figure 4] The caption in Figure 2 misspells 'Retriver' as 'Retriver'; in Figure 4 the p-values are rendered as '4.97e 54' and '8.87e 05', which should be '4.97e-54' and '8.87e-05'.
  3. [Section 4 and Table 5] The text mentions 'Claude-3.5 or GPT-4' as the LLM used for data distillation, while the Table 5 caption says 'distilled from GPT-4o'; please make the model reference consistent.
  4. [Section 3.3.4] There are typos: 'availale' appears in the text and 'not availale' in the Table 4 caption; also, the phrase 'the training set of SuperCLUE-Math6 is not availale' should be clarified as 'not available'.
  5. [Related Work] The citation 'Toshniwal et al.;' in the Related Work section lacks a year and has a stray semicolon; it should be formatted consistently with the other references.
  6. [Section 3.3.2] The description of Figure 4 as 'the top-8 retrieved data points from 100 random samples' is ambiguous: please clarify whether this means 100 random queries with 8 retrieved items each, and how the scatter plot was aggregated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central accuracy claim is an externally benchmarked comparison, and the same-label analyses are diagnostics rather than derivation steps.

full rationale

The paper's central claim is that a contrastively trained retriever using computational-graph-based positive pairs improves few-shot MWP accuracy. The retriever is trained on 25% of Math23k with positive pairs defined by G_i^+ = G_i in Eq. (1), but the reported exact-match improvements in Table 2 are computed from LLM outputs on held-out test sets of six datasets against random and BGE baselines; they are not computed from graph labels, so the main result is not forced by the label definition. The upper-bound analysis in Section 3.3.1 uses gold computational graphs directly for retrieval as an independent oracle, not as a fitted parameter, and is a diagnostic rather than a prediction. Figure 4's correlation between embedding similarity and computational-graph similarity is computed on Math23k, the same distribution used for retriever training, so it partially reflects the training objective; however, this analysis is not load-bearing for the main accuracy claim and does not define any derived quantity used in the headline results. Section 4's distilled-data experiment includes an explicit limitation that the retriever may rely on sentence-structure similarity rather than true computational graphs; this is an honest data-quality caveat, not a circular step. No load-bearing self-citations, imported uniqueness claims, or ansatz-smuggling citations appear in the derivation chain. The central result is therefore self-contained against external benchmarks.

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

The central empirical claim rests on machine-learning training choices and domain assumptions rather than a mathematical derivation. The ledger lists key chosen hyperparameters and the assumptions about computational-graph labels, embedding transfer, and answer parsing.

free parameters (3)
  • top-k number of retrieved exemplars = 8
    Chosen uniformly across all datasets and model sizes without a reported sweep; prompt effectiveness often depends on k.
  • contrastive temperature tau = 0.05
    Set for retriever training in Section 3.1; affects embedding separation and retrieval ranking.
  • retriever training data fraction = 25%
    Selected based on Figure 5 saturation analysis on one dataset, aqua_rat; not re-validated per corpus.
assumptions (4)
  • domain assumption Positive pairs defined by exact computational-graph equality capture the notion of structural analogy needed for MWP solving.
    Section 2.2 uses G+ = G as the only supervision signal; if exact graph equality is too coarse or too strict, retrieval quality is misaligned with true reasoning similarity.
  • domain assumption A text encoder trained on translated Math23k questions generalizes to other MWP datasets and languages at inference.
    Section 3.1 trains only on 25% of Math23k and uses the same retriever for ape210k, gsm8k, math_qa, Calc-ape210k, aqua_rat, and SuperCLUE-Math6; no per-dataset retraining is reported.
  • domain assumption Exact-match parsing of generated answers is a reliable measure of problem-solving correctness.
    Section 3.1 states EM accuracy with string or float matching; parsing ambiguities for multi-step text solutions could introduce measurement noise.
  • domain assumption LLM-rewritten questions preserve the intended computational graph well enough to serve as positive pairs.
    Section 4 uses Claude-3.5/GPT-4 rewrites without human verification; the authors note rewritten sentence structure is more similar than labeled pairs, which may bias training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning by Analogy: Enhancing Few-Shot Prompting for Math Word Problem Solving with Computational Graph-Based Retrieval." pith.science (2026). https://pith.science/paper/KTOHJ3AZ

@misc{pith2026241116454,
  author       = {Pith},
  title        = {Pith review of: Learning by Analogy: Enhancing Few-Shot Prompting for Math Word Problem Solving with Computational Graph-Based Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KTOHJ3AZ}},
  note         = {Machine review of arXiv:2411.16454}
}
read the original abstract

Large language models (LLMs) are known to struggle with complicated reasoning tasks such as math word problems (MWPs). In this paper, we present how analogy from similarly structured questions can improve LLMs' problem-solving capabilities for MWPs. Specifically, we rely on the retrieval of problems with similar computational graphs to the given question to serve as exemplars in the prompt, providing the correct reasoning path for the generation model to refer to. Empirical results across six math word problem datasets demonstrate the effectiveness of our proposed method, which achieves a significant improvement of up to 6.7 percent on average in absolute value, compared to baseline methods. These results highlight our method's potential in addressing the reasoning challenges in current LLMs.

Figures

Figures reproduced from arXiv: 2411.16454 by the authors.

Figure 1
Figure 1. An example of a math word problem with its [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Flowchart of Retriver Training. This figure illustrates the process of training a retriever model (encoder) with contrastive learning to identify structurally similar math word problems. Each question is encoded into an embedding based on its text. Positive pairs are formed by pairing examples with matching computational graph structures, while in-batch negatives serve as contrasting examples with different structur… view at source ↗
Figure 3
Figure 3. Case study on the retrieved data with our model and BGE respectively. The retrieved data using trained [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The scatter plot of our trained retriever (left) and BGE (right) on 100 random samples from Math23k. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: The relationship between model performance [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Some cases of the original and rewritten questions. The entity names, value of numbers and semantics are [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 13 canonical work pages

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

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

  3. [4]

    Janice Ahn, Rishu Verma, Renze Lou, Di Liu, Rui Zhang, and Wenpeng Yin. 2024. Large language models for mathematical reasoning: Progresses and challenges. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 225--237

    work page 2024

  4. [5]

    Aida Amini, Saadia Gabriel, Shanchuan Lin, Rik Koncel-Kedziorski, Yejin Choi, and Hannaneh Hajishirzi. 2019. https://doi.org/10.18653/v1/N19-1245 M ath QA : Towards interpretable math word problem solving with operation-based formalisms . In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics:...

    doi:10.18653/v1/n19-1245 2019

  5. [6]

    Swagata Ashwani, Kshiteesh Hegde, Nishith Reddy Mannuru, Mayank Jindal, Dushyant Singh Sengar, Krishna Chaitanya Rao Kathala, Dishant Banga, Vinija Jain, and Aman Chadha. 2024. https://arxiv.org/abs/2402.18139 Cause and effect: Can large language models truly understand causality? Preprint, arXiv:2402.18139

    arXiv 2024

  6. [7]

    Parishad BehnamGhader, Vaibhav Adlakha, Marius Mosbach, Dzmitry Bahdanau, Nicolas Chapados, and Siva Reddy. 2024. https://openreview.net/forum?id=IW1PR7vEBf LLM 2vec: Large language models are secretly powerful text encoders . In First Conference on Language Modeling

    2024

  7. [8]

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. 2020 a . https://arxiv.org/abs/2002.05709 A simple framework for contrastive learning of visual representations . Preprint, arXiv:2002.05709

    arXiv 2020

  8. [9]

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. 2020 b . A simple framework for contrastive learning of visual representations. In International conference on machine learning, pages 1597--1607. PMLR

    work page 2020

Show all 50 references
  1. [10]

    Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. https://arxiv.org/abs/2110.14168 Training verifiers to solve math word problems . ...

    2021 arXiv

  2. [11]

    Prakhar Dixit and Tim Oates. 2024. Sbi-rag: Enhancing math word problem solving for students through schema-based instruction and retrieval-augmented generation. arXiv preprint arXiv:2410.13293

    2024 arXiv

  3. [12]

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

    2024 arXiv

  4. [13]

    Wenqi Fan, Yujuan Ding, Liangbo Ning, Shijie Wang, Hengyun Li, Dawei Yin, Tat-Seng Chua, and Qing Li. 2024. 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,...

    2024

  5. [14]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, Meng Wang, and Haofen Wang. 2023. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997

    2023 arXiv

  6. [15]

    Gaël Gendron, Qiming Bao, Michael Witbrock, and Gillian Dobbie. 2024. https://arxiv.org/abs/2305.19555 Large language models are not strong abstract reasoners . Preprint, arXiv:2305.19555

    2024 arXiv

  7. [16]

    Rishi Hazra, Gabriele Venturato, Pedro Zuidberg Dos Martires, and Luc De Raedt. 2024. https://arxiv.org/abs/2408.07215 Can large language models reason? a characterization via 3-sat . Preprint, arXiv:2408.07215

    2024 arXiv

  8. [17]

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. 2020. Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9729--9738

    2020

  9. [18]

    Owen Henkel, Zach Levonian, Chenglu Li, and Millie Postle. 2024. Retrieval-augmented generation to improve math question-answering: Trade-offs between groundedness and human preference. In Proceedings of the 17th International Conference on Educational Data Mining, pages 315--320

    2024

  10. [19]

    Xijie Huang, Li Lyna Zhang, Kwang-Ting Cheng, and Mao Yang. 2023. Boosting llm reasoning: Push the limits of few-shot learning with reinforced in-context pruning. arXiv e-prints, pages arXiv--2312

    2023

  11. [20]

    Dan Iter, Kelvin Guu, Larry Lansing, and Dan Jurafsky. 2020. https://doi.org/10.18653/V1/2020.ACL-MAIN.439 Pretraining with contrastive sentence objectives improves discourse performance of language models . In Proceedings of the 58th Annual Meeting of the Association for Comp...

    2020 doi

  12. [21]

    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

  13. [22]

    Marek Kadlčík, Michal Štefánik, Ondřej Sotolář, and Vlastimil Martinek. 2023. https://arxiv.org/abs/2305.15017 Calc-x and calcformers: Empowering arithmetical chain-of-thought through interaction with symbolic systems . In Proceedings of the The 2023 Conference on Empirical Me...

    2023 arXiv

  14. [23]

    Yanis Labrak, Adrien Bazoge, Emmanuel Morin, Pierre-Antoine Gourraud, Mickael Rouvier, and Richard Dufour. 2024. Biomistral: A collection of open-source pretrained large language models for medical domains. arXiv preprint arXiv:2402.10373

    2024 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]

    Bangzheng Li, Ben Zhou, Fei Wang, Xingyu Fu, Dan Roth, and Muhao Chen. 2024. https://doi.org/10.18653/v1/2024.naacl-long.424 Deceptive semantic shortcuts on reasoning chains: How far can models go without hallucination? In Proceedings of the 2024 Conference of the North Americ...

    2024 doi

  17. [26]

    Jiacheng Lin, Kun Qian, Haoyu Han, Nurendra Choudhary, Tianxin Wei, Zhongruo Wang, Sahika Genc, Edward W Huang, Sheng Wang, Karthik Subbian, et al. 2024 a . Unleashing the power of llms as multi-modal encoders for text and graph-structured data. arXiv preprint arXiv:2410.11235

    2024 arXiv

  18. [27]

    Jiacheng Lin, Hanwen Xu, Zifeng Wang, Sheng Wang, and Jimeng Sun. 2024 b . Panacea: A foundation model for clinical trial search, summarization, design, and recruitment. medRxiv, pages 2024--06

    2024

  19. [28]

    Pisces: A cross-modal contrastive learning approach to synergistic drug combination prediction

    Jiacheng Lin, Hanwen Xu, Addie Woicik, and Jianzhu Ma. Pisces: A cross-modal contrastive learning approach to synergistic drug combination prediction. In Research in Computational Molecular Biology, page 268. Springer

  20. [29]

    Jiacheng Lin, Meng Xu, Zhihua Xiong, and Huangang Wang. 2024 c . https://openreview.net/forum?id=K6kt50zAiG Cambranch: Contrastive learning with augmented milps for branching . In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May...

    2024

  21. [30]

    Wang Ling, Dani Yogatama, Chris Dyer, and Phil Blunsom. 2017. Program induction by rationale generation: Learning to solve and explain algebraic word problems. ACL

    2017

  22. [31]

    Ilya Loshchilov and Frank Hutter. 2019. https://arxiv.org/abs/1711.05101 Decoupled weight decay regularization . Preprint, arXiv:1711.05101

    2019 arXiv

  23. [32]

    Eric Melz. 2023. Enhancing llm intelligence with arm-rag: Auxiliary rationale memory for retrieval augmented generation. arXiv preprint arXiv:2311.04177

    2023 arXiv

  24. [33]

    Iman Mirzadeh, Keivan Alizadeh, Hooman Shahrokhi, Oncel Tuzel, Samy Bengio, and Mehrdad Farajtabar. 2024. https://arxiv.org/abs/2410.05229 Gsm-symbolic: Understanding the limitations of mathematical reasoning in large language models . Preprint, arXiv:2410.05229

    2024 arXiv

  25. [34]

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. 2018. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748

    2018 arXiv

  26. [35]

    OpenAI, Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, et al. 2024. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . Preprint, arXiv:2303.08774

    2024 arXiv

  27. [36]

    Kv Aditya Srivatsa and Ekaterina Kochmar. 2024. What makes math word problems challenging for llms? In Findings of the Association for Computational Linguistics: NAACL 2024, pages 1138--1148

    2024

  28. [37]

    Qwen Team. 2024. https://qwenlm.github.io/blog/qwen2.5/ Qwen2.5: A party of foundation models

    2024

  29. [38]

    Openmathinstruct-2: Accelerating ai for math with massive open-source instruction data

    Shubham Toshniwal, Wei Du, Ivan Moshkov, Branislav Kisacanin, Alexan Ayrapetyan, and Igor Gitman. Openmathinstruct-2: Accelerating ai for math with massive open-source instruction data. In The 4th Workshop on Mathematical Reasoning and AI at NeurIPS'24

  30. [39]

    Liang Wang, Nan Yang, Xiaolong Huang, Binxing Jiao, Linjun Yang, Daxin Jiang, Rangan Majumder, and Furu Wei. 2022. Text embeddings by weakly-supervised contrastive pre-training. arXiv preprint arXiv:2212.03533

    2022 arXiv

  31. [40]

    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

  32. [41]

    Yan Wang, Xiaojiang Liu, and Shuming Shi. 2017. https://doi.org/10.18653/v1/D17-1088 Deep neural solver for math word problems . In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 845--854, Copenhagen, Denmark. Association for Comp...

    2017 doi

  33. [42]

    Chi, Tatsunori Hashimoto, Oriol Vinyals, Percy Liang, Jeff Dean, and William Fedus

    Jason Wei, Yi Tay, Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgeaud, Dani Yogatama, Maarten Bosma, Denny Zhou, Donald Metzler, Ed H. Chi, Tatsunori Hashimoto, Oriol Vinyals, Percy Liang, Jeff Dean, and William Fedus. 2022 a . https://arxiv.org/abs/2206.07682 Emer...

    2022 arXiv

  34. [43]

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

    2022

  35. [44]

    Shitao Xiao, Zheng Liu, Peitian Zhang, and Niklas Muennighoff. 2023. https://arxiv.org/abs/2309.07597 C-pack: Packaged resources to advance general chinese embedding . Preprint, arXiv:2309.07597

    2023 arXiv

  36. [45]

    H Xu, J Lin, A Woicik, Z Liu, J Ma, S Zhang, H Poon, L Wang, and S Wang. 2022. Pisces: A multi-modal data augmentation approach for drug combination synergy prediction

    2022

  37. [46]

    Liang Xu, Hang Xue, Lei Zhu, and Kangkang Zhao. 2024. https://arxiv.org/abs/2401.11819 Superclue-math6: Graded multi-step math reasoning benchmark for llms in chinese . Preprint, arXiv:2401.11819

    2024 arXiv

  38. [47]

    An Yang, Beichen Zhang, Binyuan Hui, Bofei Gao, Bowen Yu, Chengpeng Li, Dayiheng Liu, Jianhong Tu, Jingren Zhou, Junyang Lin, et al. 2024. Qwen2. 5-math technical report: Toward mathematical expert model via self-improvement. arXiv preprint arXiv:2409.12122

    2024 arXiv

  39. [48]

    Junliang Yu, Hongzhi Yin, Xin Xia, Tong Chen, Lizhen Cui, and Quoc Viet Hung Nguyen. 2022. Are graph augmentations necessary? simple graph contrastive learning for recommendation. In Proceedings of the 45th international ACM SIGIR conference on research and development in info...

    2022

  40. [49]

    Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu

    Longhui Yu, Weisen Jiang, Han Shi, Jincheng Yu, Zhengying Liu, Yu Zhang, James T. Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. 2024 a . https://openreview.net/forum?id=N8N0hgNDRt Metamath: Bootstrap your own mathematical questions for large language models . In The Twelft...

    2024

  41. [50]

    Ping Yu, Jing Xu, Jason Weston, and Ilia Kulikov. 2024 b . https://arxiv.org/abs/2407.06023 Distilling system 2 into system 1 . Preprint, arXiv:2407.06023

    2024 arXiv

  42. [51]

    Wei Zhao, Mingyue Shang, Yang Liu, Liang Wang, and Jingming Liu. 2020. https://arxiv.org/abs/2009.11506 Ape210k: A large-scale and template-rich dataset of math word problems . Preprint, arXiv:2009.11506

    2020 arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.