Pith. sign in

REVIEW 4 major objections 6 minor 37 references

Learn to Think: Bootstrapping LLM Reasoning Capability Through Graph Representation Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read L2T claims that representing a model's reasoning as a graph and training a small GNN controller lets one LLM outperform CoT, ToT, GoT, and AoT across puzzles and writing tasks without task-specific prompts.

desk verdict L2T's graph-controlled reasoning is a real step forward, but the paper must disclose its RL training split before the 'no additional training' claim can stand. read the letter →

arxiv 2505.06321 v2 pith:KFCC6XRE submitted 2025-05-09 cs.LG cs.AI

Hang Gao , Chenhao Zhang , Tie Wang , Junsuo Zhao , Fengge Wu , Changwen Zheng , Huaping Liu This is my paper · ORCID
classification cs.LGcs.AI
keywords largelanguagemodelsreasoninggraphrepresentationlearningneuralnetworkschain-of-thoughtreinforcementpromptengineeringadaptive
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 introduces Learn to Think (L2T), a framework whose central claim is that a large language model can reason across many different tasks using one generic prompt, provided its reasoning process is represented as a graph and steered by a learned controller. Each thought becomes a node; the LLM labels open nodes as stop, continue, output, or backtrack, and a small graph neural network decides how the next thoughts are generated. L2T is evaluated on Sudoku, Game of 24, TruthQuest, and creative-writing benchmarks and reported to outperform CoT, ToT, GoT, and AoT, especially when task-specific prompts are removed. If correct, the framework points to a general reasoning controller that transfers across tasks without fine-tuning the underlying model.

What carries the argument

The central object is the reasoning process graph, a directed graph whose nodes are the LLM's thoughts and whose edges point from a thought to its successor, with each pending node assigned one of four labels by the LLM itself: stop, continue, final output, or backtrack. The graph unifies chain-, tree-, and graph-style reasoning because those are special cases. The load-bearing mechanism is the loop between the LLM classifier and a GNN controller: the LLM labels pending nodes from an induced subgraph rendered as text (Eq. 2), the controller's one-layer GCN plus MLP maps the graph to an action vector controlling branch count, temperature, and top-p, thought generation follows that action (Eq. 4), and PPO updates the controller using the LLM's evaluation score as reward. The graph state is what lets the controller see the whole reasoning process before choosing the next move.

What would settle it

Take a sample of reasoning steps from the Sudoku and Game of 24 runs, have a human oracle label each pending node as stop, continue, final result, or backtrack, and compare with the LLM's assigned labels; then check whether the LLM's reward scores predict which reasoning threads reach the correct final answer. If label agreement is near chance or reward scores are uncorrelated with ultimate success, the bootstrap has no training signal and the reported gains would not replicate.

Watch

Extended reading notes

Core claim

The paper claims that LLM reasoning can be bootstrapped without task-specific prompt design by making the reasoning process itself the learning target. Every thought the model produces is a node in a directed reasoning graph; the LLM labels each unprocessed node with one of four classes—stop, continue, output as final result, or backtrack—and only 'continue' nodes generate children. A one-layer graph convolutional network reads the current graph and outputs an adjustment vector that sets the number of branches, temperature, and top-p for the next thought, and this controller is trained with PPO using an integer LLM self-score as reward. In the reported experiments L2T outperforms CoT, ToT, GoT, and AoT on all four evaluated tasks, including the variants where task-specific prompts are stripped from the baselines, and the no-GNN variant also beats the baselines, supporting the claim that the graph-level labels carry most of the benefit.

Load-bearing premise

The method stands on the assumption that the same LLM can reliably label its own intermediate thoughts with the four node classes and assign trustworthy 0–10 reward scores from a generic evaluation prompt; the paper does not measure either reliability directly.

Editorial extensions

If this is right

  • A single generic prompt is enough for heterogeneous tasks such as Sudoku, arithmetic puzzles, logic puzzles, and creative writing, so users can stop hand-designing task-specific reasoning prompts.
  • Removing task-specific prompts sharply degrades CoT, ToT, GoT, and AoT, while L2T, which never uses them, keeps its reported full-strength performance.
  • The learned controller uses fewer reasoning steps: L2T generates fewer graph nodes than L2T without the GNN module, meaning the policy makes the model think more efficiently, not just more accurately.
  • Reinforcement learning on the LLM's own evaluation scores improves the controller beyond direct score-based training, as shown by the L2T versus L2T w/o RL ablation.
  • The controller adapts its strategy to the task, choosing correlated temperature/top-p values for the arithmetic task and anti-correlated values for creative writing.

Reading between the lines

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

  • Because the controller acts on a task-agnostic graph state, the same trained policy could be applied to other generative agent loops—tool use, multi-turn dialogue, or planning—where each action can be represented as a node.
  • The reliance on LLM-assigned labels and reward scores suggests a direct test: if a different, weaker base LLM is swapped in, the controller's performance will reveal how much of the gain comes from the graph policy versus the base model's own grading ability.
  • The action vector could be extended beyond sampling parameters to choices like which subgraph to revisit or which tool to call, which would make 'reasoning mode' a more general action space.
  • A zero-shot test on a novel task type not seen in training would tell whether the controller has learned to allocate exploration and exploitation in a transferable way rather than memorizing task-specific patterns.
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.

Referee Report

4 major / 6 minor

Summary. The paper proposes L2T, a framework that represents an LLM's step-by-step reasoning as a graph, uses the LLM itself to classify each thought into one of four states (stop, continue, complete, backtrack), and introduces a GNN-based module that selects reasoning modes (branch count, temperature, top-p) during inference. The GNN module is trained with PPO. The authors claim that L2T improves reasoning performance across Sudoku, Game of 24, TruthQuest, and Creative Writing without task-specific prompts and without additional LLM training, and they compare against IO, CoT, ToT, GoT, and AoT baselines. The paper also reports ablations (L2T w/o GNN, L2T w/o RL, L2T w/ MLP) and token-consumption analyses.

Significance. If the claims hold, the method is a meaningful contribution: it offers a unified reasoning scaffold that adapts its search strategy online, and the objective results on Sudoku and Game of 24 show large improvements over structured prompting baselines. The inclusion of the L2T w/o GNN ablation is a strong feature, because it isolates the contribution of the graph-based thought-management mechanism from the learned GNN module. The paper also releases code and provides token-cost comparisons. However, the headline claim of working 'without additional training' is not supported as stated, because the GNN controller is trained with PPO, and the paper does not disclose the training/evaluation split that would establish task-agnostic behavior. The Creative Writing evaluation is also incomplete because Table 4 omits the L2T row.

major comments (4)
  1. [Abstract; Section 3.3; Appendix A.2] The abstract claims that L2T improves reasoning 'without requiring additional training or task-specific prompt design,' but Section 3.3 and Appendix A.2 describe a PPO-trained GNN-based controller trained over 20 epochs. The paper never states which problems or data generated the reinforcement-learning trajectories. If those trajectories come from the same Sudoku, Game of 24, TruthQuest, or Creative Writing instances that are later reported in Tables 1-4, then the controller is task-specifically optimized and the comparison against zero-shot and few-shot baselines does not test a task-agnostic method. The authors must disclose the exact training/evaluation split and clarify in the abstract and intro that 'no additional training' refers only to the LLM, not to the GNN controller.
  2. [Table 4; Section 4.1 Creative Writing results] Table 4 omits the L2T row even though every percentage is defined relative to L2T ('Higher'/'Same'/'Lower'). Consequently, the reader cannot directly verify the text's claim that 'L2T achieves higher or equivalent scores in over 80% of cases' for the comparison against L2T w/o GNN; for the Sentence Formation (Less Hints) column, L2T is lower than L2T w/o GNN in 34.11% of cases and higher in only 15.05%. This pattern is not discussed and appears to contradict the paper's implication that the GNN module improves open-ended generation quality. Please add the L2T row (or an equivalent reference) and explain the relationship between L2T and L2T w/o GNN on Creative Writing.
  3. [Section 3.2 Eq. (2); Section 3.3 reward r_k] The four-way node classification that drives the whole reasoning process is performed by the same GPT-4o model that generates the thoughts, and the reward r_k used to train the GNN is an integer from 0 to 10 assigned by that same LLM based on the LLM-generated evaluation criteria Xeva. The paper never measures the classification accuracy of Y_v against any ground truth, nor does it measure how well r_k correlates with final task success. If the node classifier is unreliable, or if r_k rewards the LLM's stylistic preferences rather than correctness, the PPO-trained controller could be optimizing a proxy that is disconnected from the objective task metrics. The authors should provide a human- or gold-labeled evaluation of the node classifier and a correlation analysis between r_k and final correctness on at least one task.
  4. [Section 4.1 Results; Tables 1-3] The text repeatedly uses the phrase 'significant improvements,' but no statistical significance tests are reported anywhere in the paper. Tables 1-3 report means and standard deviations, and some differences are large (for example, 89.23 versus 69.41 on 5x5 Sudoku), but the claim of significance is not supported by any paired test or confidence interval. Given that the Sudoku results use only 13 puzzle sets per condition, the authors should either perform an appropriate paired test (e.g., Wilcoxon signed-rank) or soften the wording to 'consistent improvements' with the reported variability.
minor comments (6)
  1. [Introduction, first paragraph] Tree of Thoughts is cited to [Chu et al., 2024], but the correct reference for ToT is [Yao et al., 2023a], which is cited later in the Related Works section.
  2. [Figure 1] The bar chart in Figure 1 contains untranslated placeholder labels such as '类别 1' and '系列4'; these should be replaced with the intended English axis labels and legend entries.
  3. [Section 3.2, Eq. (1)] The set-builder notation in Eq. (1) is missing a set-membership symbol; it should read $\tilde{E}_v^{(k-1)} = \{(u,w) \in E^{(k-1)} \mid u \in \tilde{V}_v^{(k-1)}, w \in \tilde{V}_v^{(k-1)}\}$.
  4. [Section 4.1, Tasks; Appendix B.4] The Creative Writing task is described as 'a series of diverse writing challenges,' but Appendix B.4 describes only two specific tasks, and the mapping between those two tasks and the three columns of Table 4 (Sentence Formation with fewer hints, Sentence Formation with more hints, Text Expansion) is not explained.
  5. [Table 4 caption] The caption phrase 'All data represent the performance of L2T comparisons to other methods' is awkward; it would be clearer to state that each row gives the percentage of cases in which L2T scores higher, the same, or lower than the named method.
  6. [Appendix A.5, Format Generation Prompt] The phrasing 'Each solution must consist of at least two or more steps' mixes two comparatives; it should be 'at least two steps' or 'two or more steps.'

Circularity Check

2 steps flagged · score 4.0 of 10

Self-referential LLM reward and LLM-judged Creative Writing scores create a partial circularity loop; external Sudoku, Game of 24, and TruthQuest results remain independently grounded.

  1. self definitional [Section 3.3 (reward r_k and PPO update); Appendix A.5 (Evaluation Information Generation Prompt and Evaluation Prompt); Figure 3 caption]
    "The reward rk is set to 100 if the generated thought represents the final result. Otherwise, it is an integer between 0 and 10, determined by the LLM based on G(k) and Xeva. ... All LLM modules uniformly utilize the same LLM."

    The GNN-based reasoning mode selection module is trained with PPO using r_k as the reward, and Xeva is generated by the same LLM from the task description. The same LLM then scores each generated thought, so the policy is optimized to maximize the LLM's self-assessment of its own reasoning. The paper's ablation claim that the GNN module contributes (Table 5) is therefore an improvement against a self-generated reward, and the paper never measures the reliability of that reward or of the LLM's node classifications. The loop is closed by construction: the controller is fit to the same LLM judgment that defines the intermediate quality signal.

  2. self definitional [Section 4.1 (Table 4 and surrounding text); Figure 3 caption]
    "Table 4 presents results on Creative Writing, focusing on relative scores to L2T. Evaluations via an LLM reduce fluctuations. ... All LLM modules uniformly utilize the same LLM."

    For Creative Writing, no human or external rubric is reported; the only scores are produced by an LLM. The same LLM generates the thoughts, extracts the evaluation criteria Xeva, and supplies the reward used to train the GNN controller. Consequently, the reported 'higher or equivalent in over 80% of cases' on Creative Writing is the trained system being graded by the very scoring function it was optimized to maximize. This makes the open-ended-task result a self-consistent loop rather than independent evidence. The missing L2T row in Table 4 further prevents the reader from checking the actual scores, though that omission is a reporting issue rather than the circularity itself.

full rationale

The three constrained tasks (Sudoku, Game of 24, TruthQuest) are checked against exact, externally verifiable answers in Tables 1-3, so the central claim of improved reasoning on those tasks does not reduce to a fit or to a self-citation chain. The paper contains no load-bearing self-citations and no imported uniqueness theorem. The concrete circularity is localized to the GNN training reward and the Creative Writing evaluation: both use the same LLM that is being guided, with evaluation criteria Xeva generated by that same LLM. This is a partial, not total, circularity because the main benchmark results remain externally grounded. Separately, the abstract's 'without requiring additional training' claim conflicts with the 20-epoch PPO training stated in Appendix A.2, and the missing training/evaluation split for the RL trajectories is an unverified experimental-design risk; I treat those as correctness and reporting concerns rather than as demonstrated circular steps.

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

The method's claims rest on empirical behavior of GPT-4o as a classifier and judge, not on closed-form derivations, so the ledger is dominated by domain assumptions about LLM reliability.

free parameters (5)
  • context window beta = 2
    Predefined hyperparameter in Section A.2; limits how many prior thoughts are included in the textual subgraph passed to the LLM for node classification.
  • PPO learning rate = 5e-3
    Set in Section A.2 for training the GNN policy and value MLPs.
  • PPO training epochs = 20
    Set in Section A.2 for reinforcement learning of the mode-selection module.
  • PPO clip parameter = 0.2
    Standard PPO clipping threshold, Section A.2.
  • Reward for non-final thoughts = integer 0-10 via LLM
    Reward design in Section 3.3; the LLM scores each thought against auto-generated criteria X_eva.
assumptions (4)
  • domain assumption GPT-4o's classifications of reasoning nodes (labels 1-4) are accurate enough to guide search
    Section 3.2 defines node classification via Eq. 2; no accuracy measurement is reported.
  • domain assumption LLM-generated criteria X_eva and the LLM's 0-10 scores are valid rewards for training the controller
    Section 3.3 sets rk from the LLM based on G(k) and X_eva; validity is untested.
  • domain assumption One-layer GCN on final-token embeddings of thoughts provides sufficient state for reasoning-mode selection
    Section 3.2 uses h_v from the LLM's last output token as node features; no comparison to deeper encoders.
  • domain assumption PPO with the stated hyperparameters converges to a useful policy for this small controller
    Section A.2 gives 20 epochs, LR 5e-3; no convergence analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learn to Think: Bootstrapping LLM Reasoning Capability Through Graph Representation Learning." pith.science (2026). https://pith.science/paper/KFCC6XRE

@misc{pith2026250506321,
  author       = {Pith},
  title        = {Pith review of: Learn to Think: Bootstrapping LLM Reasoning Capability Through Graph Representation Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KFCC6XRE}},
  note         = {Machine review of arXiv:2505.06321}
}
read the original abstract

Large Language Models (LLMs) have achieved remarkable success across various domains. However, they still face significant challenges, including high computational costs for training and limitations in solving complex reasoning problems. Although existing methods have extended the reasoning capabilities of LLMs through structured paradigms, these approaches often rely on task-specific prompts and predefined reasoning processes, which constrain their flexibility and generalizability. To address these limitations, we propose a novel framework that leverages graph learning to enable more flexible and adaptive reasoning capabilities for LLMs. Specifically, this approach models the reasoning process of a problem as a graph and employs LLM-based graph learning to guide the adaptive generation of each reasoning step. To further enhance the adaptability of the model, we introduce a Graph Neural Network (GNN) module to perform representation learning on the generated reasoning process, enabling real-time adjustments to both the model and the prompt. Experimental results demonstrate that this method significantly improves reasoning performance across multiple tasks without requiring additional training or task-specific prompt design. Code can be found in https://github.com/zch65458525/L2T.

Figures

Figures reproduced from arXiv: 2505.06321 by the authors.

Figure 1
Figure 1. A comparison between our method and conventional [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An example of the reasoning process graph. Each box contains a thought generated by the LLM, representing a node in the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The framework of the proposed method. All LLM modules uniformly utilize the same LLM. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The temperature and top-p value within the reasoning pro￾cess. Acknowledgments We would like to express our sincere gratitude to the review￾ers of this paper, as well as the Program Committee and Area Chairs, for their valuable comments and suggestions. This work is su…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 22 canonical work pages

  1. [1]

    Graph of thoughts: Solving elaborate problems with large language models

    Maciej Besta, Nils Blach, Ales Kubicek, Robert Gerstenberger, Michal Podstawski, Lukas Gianinazzi, Joanna Gajda, Tomasz Lehmann, Hubert Niewiadomski, Piotr Nyczyk, and Torsten Hoefler. Graph of thoughts: Solving elaborate problems with large language models. In Michael J. Wooldridge, Jennifer G. Dy, and Sriraam Natarajan, editors, Thirty-Eighth AAAI Confe...

    work page 2024

  2. [2]

    Luca Beurer - Kellner, Mark Niklas M \" u ller, Marc Fischer, and Martin T. Vechev. Prompt sketching for large language models. In Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-27, 2024 . OpenReview.net, 2024

    work page 2024

  3. [3]

    Language models are few-shot learners

    Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in Neural Information Processing Systems , 2022

    work page 2022

  4. [4]

    Navigate through enigmatic labyrinth A survey of chain of thought reasoning: Advances, frontiers and future

    Zheng Chu, Jingchang Chen, Qianglong Chen, Weijiang Yu, Tao He, Haotian Wang, Weihua Peng, Ming Liu, Bing Qin, and Ting Liu. Navigate through enigmatic labyrinth A survey of chain of thought reasoning: Advances, frontiers and future. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 202...

    work page 2024

  5. [5]

    Chain-of-verification reduces hallucination in large language models

    Shehzaad Dhuliawala, Mojtaba Komeili, Jing Xu, Roberta Raileanu, Xian Li, Asli Celikyilmaz, and Jason Weston. Chain-of-verification reduces hallucination in large language models. In Lun - Wei Ku, Andre Martins, and Vivek Srikumar, editors, Findings of the Association for Computational Linguistics, ACL 2024, Bangkok, Thailand and virtual meeting, August 1...

    work page 2024

  6. [6]

    Active prompting with chain-of-thought for large language models

    Shizhe Diao, Pengcheng Wang, Yong Lin, Rui Pan, Xiang Liu, and Tong Zhang. Active prompting with chain-of-thought for large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok, Thailand, August 11-16, 2024 , pages 1330--1350. Association for Computational Li...

    work page 2024

  7. [7]

    Large language models: a comprehensive survey of its applications, challenges, limitations, and future prospects

    Muhammad Usman Hadi, Qasem Al Tashi, Abbas Shah, Rizwan Qureshi, Amgad Muneer, Muhammad Irfan, Anas Zafar, Muhammad Bilal Shaikh, Naveed Akhtar, Jia Wu, et al. Large language models: a comprehensive survey of its applications, challenges, limitations, and future prospects. Authorea Preprints , 2024

    work page 2024

  8. [8]

    Recent advances in neural machine translation

    Wei Jiao, Yingce Xia, Tao Qin, Nenghai Yu, and Tie-Yan Liu. Recent advances in neural machine translation. AI Open , 3:36--45, 2022

    work page 2022

Show all 37 references
  1. [9]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings . OpenReview.net, 2017

    2017

  2. [10]

    Konda and John N

    Vijay R. Konda and John N. Tsitsiklis. Actor-critic algorithms. In Sara A. Solla, Todd K. Leen, and Klaus - Robert M \" u ller, editors, Advances in Neural Information Processing Systems 12, [NIPS Conference, Denver, Colorado, USA, November 29 - December 4, 1999] , pages 1008-...

    1999

  3. [11]

    u ttler, Mike Lewis, Wen - tau Yih, Tim Rockt \

    Patrick S. H. Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \" u ttler, Mike Lewis, Wen - tau Yih, Tim Rockt \" a schel, Sebastian Riedel, and Douwe Kiela. Retrieval-augmented generation for knowledge-intensive NLP tasks. In ...

    2020

  4. [12]

    Large language models understand and can be enhanced by emotional stimuli, 2023

    Cheng Li, Jindong Wang, Yixuan Zhang, Kaijie Zhu, Wenxin Hou, Jianxun Lian, Fang Luo, Qiang Yang, and Xing Xie. Large language models understand and can be enhanced by emotional stimuli, 2023

    2023

  5. [13]

    Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing

    Pengfei Liu, Weizhe Yuan, Jinlan Fu, Zhengbao Jiang, Hiroaki Hayashi, and Graham Neubig. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Computing Surveys , 55(9):1--35, 2023

    2023

  6. [14]

    Large language model guided tree-of-thought

    Jieyi Long. Large language model guided tree-of-thought. arXiv preprint arXiv:2305.08291 , 2023

    2023 arXiv

  7. [15]

    Directgpt: A direct manipulation interface to interact with large language models

    Damien Masson, Sylvain Malacria, G \' e ry Casiez, and Daniel Vogel. Directgpt: A direct manipulation interface to interact with large language models. In Florian 'Floyd' Mueller, Penny Kyburz, Julie R. Williamson, Corina Sas, Max L. Wilson, Phoebe O. Toups Dugas, and Irina Sh...

    2024

  8. [16]

    Liar, liar, logical mire: A benchmark for suppositional reasoning in large language models

    Philipp Mondorf and Barbara Plank. Liar, liar, logical mire: A benchmark for suppositional reasoning in large language models. In Yaser Al - Onaizan, Mohit Bansal, and Yun - Nung Chen, editors, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Process...

    2024

  9. [17]

    Codegen: An open large language model for code with multi-turn program synthesis

    Vincent Ni, Adrian Lee, Shruti Kumar, Saikrishna Chalamalasetti, Aditi Singh, Nadjet Tazi, Dhruva Patil, et al. Codegen: An open large language model for code with multi-turn program synthesis. arXiv preprint arXiv:2203.13474 , 2022

    2022 arXiv

  10. [18]

    GPT-4 technical report

    OpenAI. GPT-4 technical report. CoRR , abs/2303.08774, 2023

    2023 arXiv

  11. [19]

    Lundberg, Sameer Singh, Hannaneh Hajishirzi, Luke Zettlemoyer, and Marco T \' u lio Ribeiro

    Bhargavi Paranjape, Scott M. Lundberg, Sameer Singh, Hannaneh Hajishirzi, Luke Zettlemoyer, and Marco T \' u lio Ribeiro. ART: automatic multi-step reasoning and tool-use for large language models. CoRR , abs/2303.09014, 2023

    2023 arXiv

  12. [20]

    Carbon emissions and large neural network training

    David Patterson, Joseph Gonzalez, Quoc Le, Chen Liang, Lluis Munguia, Daniel Rothchild, David So, Marc Texier, and Jeff Dean. Carbon emissions and large neural network training. arXiv preprint arXiv:2204.05149 , 2022

    2022 arXiv

  13. [21]

    Reasoning with language model prompting: A survey

    Shuofei Qiao, Yixin Ou, Ningyu Zhang, Xiang Chen, Yunzhi Yao, Shumin Deng, Chuanqi Tan, Fei Huang, and Huajun Chen. Reasoning with language model prompting: A survey. In Anna Rogers, Jordan L. Boyd - Graber, and Naoaki Okazaki, editors, Proceedings of the 61st Annual Meeting o...

    2023

  14. [22]

    Improving language understanding by generative pre-training

    Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. Improving language understanding by generative pre-training. OpenAI Blog , 1(8):1--12, 2018

    2018

  15. [23]

    Proximal policy optimization algorithms

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms. CoRR , abs/1707.06347, 2017

    2017 arXiv

  16. [24]

    Algorithm of thoughts: Enhancing exploration of ideas in large language models

    Bilgehan Sel, Ahmad Al - Tawaha, Vanshaj Khattar, Ruoxi Jia, and Ming Jin. Algorithm of thoughts: Enhancing exploration of ideas in large language models. In Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-27, 2024 . OpenReview.net, 2024

    2024

  17. [25]

    Black-box tuning for language-model-as-a-service

    Kaiyuan Sun, Cheng Zhou, Deng Cai, and Ming Ding. Black-box tuning for language-model-as-a-service. arXiv preprint arXiv:2201.03514 , 2022

    2022 arXiv

  18. [26]

    Le, Ed H

    Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc V. Le, Ed H. Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. Self-consistency improves chain of thought reasoning in language models. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali,...

    2023

  19. [27]

    Chi, Quoc V

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed H. Chi, Quoc V. Le, and Denny Zhou. Chain-of-thought prompting elicits reasoning in large language models. In Sanmi Koyejo, S. Mohamed, A. Agarwal, Danielle Belgrave, K. Cho, and A. Oh, editors, ...

    2022

  20. [28]

    A comprehensive survey on graph neural networks

    Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and S Yu Philip. A comprehensive survey on graph neural networks. IEEE transactions on neural networks and learning systems , 32(1):4--24, 2020

    2020

  21. [29]

    Tree of thoughts: Deliberate problem solving with large language models

    Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Tom Griffiths, Yuan Cao, and Karthik Narasimhan. Tree of thoughts: Deliberate problem solving with large language models. In Alice Oh, Tristan Naumann, Amir Globerson, Kate Saenko, Moritz Hardt, and Sergey Levine, editors, Adva...

    2023

  22. [30]

    Narasimhan, and Yuan Cao

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R. Narasimhan, and Yuan Cao. React: Synergizing reasoning and acting in language models. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023 . OpenRevie...

    2023

  23. [31]

    Chain-of-note: Enhancing robustness in retrieval-augmented language models

    Wenhao Yu, Hongming Zhang, Xiaoman Pan, Kaixin Ma, Hongwei Wang, and Dong Yu. Chain-of-note: Enhancing robustness in retrieval-augmented language models. CoRR , abs/2311.09210, 2023

    2023 arXiv

  24. [32]

    Automatic chain of thought prompting in large language models

    Zhuosheng Zhang, Aston Zhang, Mu Li, and Alex Smola. Automatic chain of thought prompting in large language models. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023 . OpenReview.net, 2023

    2023

  25. [33]

    Enhancing zero-shot chain-of-thought reasoning in large language models through logic

    Xufeng Zhao, Mengdi Li, Wenhao Lu, Cornelius Weber, Jae Hee Lee, Kun Chu, and Stefan Wermter. Enhancing zero-shot chain-of-thought reasoning in large language models through logic. In Nicoletta Calzolari, Min - Yen Kan, V \' e ronique Hoste, Alessandro Lenci, Sakriani Sakti, a...

    2024

  26. [34]

    Learning to prompt for vision-language models

    Kevin Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models. International Journal of Computer Vision , 130(9):2337--2348, 2022

    2022

  27. [35]

    Large language models are human-level prompt engineers

    Yongchao Zhou, Andrei Ioan Muresanu, Ziwen Han, Keiran Paster, Silviu Pitis, Harris Chan, and Jimmy Ba. Large language models are human-level prompt engineers. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023 . Open...

    2023

  28. [36]

    Thread of thought unraveling chaotic contexts

    Yucheng Zhou, Xiubo Geng, Tao Shen, Chongyang Tao, Guodong Long, Jian - Guang Lou, and Jianbing Shen. Thread of thought unraveling chaotic contexts. CoRR , abs/2311.08734, 2023

    2023 arXiv

  29. [37]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence '...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.