Pith. sign in

REVIEW 4 major objections 4 minor 2 cited by

Enhancing Function-Calling Capabilities in LLMs: Strategies for Prompt Formats, Data Integration, and Multilingual Translation

T0 review · 4 major / 4 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 This paper argues that instruction-following data, when mixed into function-calling fine-tuning, substantially improves both the structural accuracy of generated tool calls and the model's ability to detect when no tool is relevant.

desk verdict Useful recipe paper on function-calling fine-tuning, but its headline claim about instruction-following data is confounded with dataset size, and the Decision Token's benefit actually comes from synthetic negative data. read the letter →

arxiv 2412.01130 v2 pith:WS7E3VKD submitted 2024-12-02 cs.CL

Yi-Chang Chen , Po-Chun Hsu , Chan-Jan Hsu , Da-shan Shiu This is my paper · ORCID
classification cs.CL
keywords functioncallingdecisiontokeninstruction-followingdatarelevancedetectionmultilingualtranslationchain-of-thoughtreasoningfine-tuningsynthetic
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 tries to show that the way you format function descriptions and, more importantly, the data you mix into fine-tuning determine how well a large language model can call tools. Its central claim is that instruction-following data is not a distraction: adding 110k instruction-following examples to 110k function-calling examples raises function-call structural accuracy from 74.62 to 85.25 on the Berkeley Function Calling Leaderboard and raises relevance detection from 38.33 to 49.58. It also proposes a Decision Token, a pair of special output tokens the model emits before answering or calling a tool, which makes it easy to create synthetic 'no tool needed' training data and improves the model's ability to refrain from calling irrelevant tools. A dedicated translation pipeline for converting English function-calling data into Traditional Chinese is shown to lift a Traditional Chinese benchmark substantially. If true, these results say that reliable tool use is more about training-data composition and a clean decision point than about prompt format alone.

What carries the argument

The central mechanism is the Decision Token: a pair of special tokens, <|answer|> and <|use_tool|>, that the model is trained to output first before generating a direct response or a tool call. Because next-token prediction is a classification, the Decision Token turns the 'should I call a tool?' judgement into an explicit binary classification. It also enables a cheap data-generation trick: take a sample where function A was called, delete A and keep functions B and C, assume B and C are irrelevant, and label the sample as an answer-directly case; training on such synthetic non-function-call data teaches the model to refrain from hallucinated calls.

What would settle it

Take a set of function-call samples, delete the called function, but leave in place one remaining function that is a near-duplicate (same purpose, different name). Fine-tune with the Decision Token + NF-1k recipe and test on relevance detection; if the model does not beat the baseline on this test, the synthetic-label assumption fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the core discovery is that mixing instruction-following data into function-calling fine-tuning significantly improves both function-calling accuracy and relevance detection, and that a learned decision token combined with synthetic non-function-call data further improves the model's ability to know when not to call any tool. The evidence is a set of fine-tuning experiments on Breeze-7B evaluated on the Berkeley Function Calling Leaderboard and a Traditional Chinese translation of it.

Load-bearing premise

The synthetic non-function-call data labels assume that after deleting the helpful function, the remaining functions are truly irrelevant to the user query, so the 'answer directly' label is correct.

Editorial extensions

If this is right

  • Instruction-following data should be considered an essential ingredient of function-calling fine-tuning, not a separate capability to be trained in a separate stage.
  • Putting function descriptions in a dedicated role, rather than mixing them into the system prompt, makes relevance detection easier to learn (49.58 vs 39.58 on the paper's baseline).
  • The Decision Token recipe can be used to generate non-function-call data from any existing function-call dataset at scale, without needing hand-written negative responses.
  • A translation pipeline that preserves function names and only translates arguments where sensible can extend function-calling models to a new language with a relatively small amount (19k examples) of translated data.
  • The translation pipeline also improves Traditional Chinese relevance detection, not just structural calling accuracy.

Reading between the lines

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

  • An implication the authors leave implicit is that the same synthetic-negative trick generalizes to any slot that can be framed as a binary decision (e.g., whether to retrieve, whether to escalate), not just tool calling.
  • The instruction-following benefit may come from the model learning prompt-semantic structure rather than from the content of the instructions; a testable extension would be to mix in instruction data with tool-related vocabulary removed and see if the gain persists.
  • The Decision Token could serve as a cheap confidence signal: the probability assigned to <|use_tool|> versus <|answer|> might be used at inference time to decide when to defer to a fallback, a use the paper does not explore.
  • The paper's Traditional Chinese gain suggests the biggest multilingual bottleneck is the format of the data rather than the language itself, which would predict that the same pipeline works for other low-resource languages with minimal adaptation.
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 / 4 minor

Summary. The paper studies how to improve the function-calling abilities of an LLM by varying prompt formats, mixing instruction-following data with function-calling data, introducing a pair of special Decision Tokens, adding synthetic non-function-call data, incorporating chain-of-thought reasoning, and using a translation pipeline for Traditional Chinese. The authors fine-tune Breeze-7B with LoRA on datasets IF-110k and FC-110k, evaluate on BFCL, MT-Bench, and a Traditional Chinese function-calling benchmark, and report three main claims: instruction-following data improves function-calling accuracy and relevance detection; the Decision Token combined with synthetic non-function-call data improves relevance detection; and the translation pipeline improves Traditional Chinese function calling. The experimental results are given in Tables 1-3.

Significance. If the claims were fully established, the paper would provide useful engineering guidance for practitioners building function-calling models, particularly regarding data mixing and relevance detection. However, the current evidence is weakened by confounded comparisons, missing ablations, lack of statistical rigor, and the use of a benchmark built in part by the authors. The contribution is therefore more of a promising empirical direction than a definitive conclusion.

major comments (4)
  1. [4.2, Table 1] The central claim that instruction-following data improves function-calling accuracy and relevance detection is not supported by the reported design. In Table 1, rows (b) versus (d) and (c) versus (e) compare FC-110k alone with FC-110k plus IF-110k, so the treatment manifoldly changes both the data content and the total training corpus size (110k vs. 220k examples). Without a quantity-matched or content-matched control, such as FC-220k without IF data or an additional 110k generic chat examples, the observed gains in AST Summary and Relevance Detection cannot be attributed specifically to instruction-following content; they could result from simple data-volume increases, regularization, or greater prompt diversity. Please add a control that isolates the effect of the IF data and, preferably, report multiple seeds.
  2. [4.3, Table 2] The Decision Token claim is underdetermined because no ablation isolates the token from the synthetic non-function-call data. In Table 2, adding the Decision Token alone actually lowers Relevance Detection in the dedicated-role condition (49.58 to 37.50) and raises it in the system-role condition (39.58 to 47.50), while the large increases appear only after NF-1k is added (57.50 and 65.42). Since no NF-only condition is reported, it is possible that the synthetic non-function-call data, not the Decision Token, drives the benefit. Please run an NF-only ablation and report all intermediate configurations.
  3. [3.2] The synthetic non-function-call data generation relies on the assumption that after removing the original helpful function func_A, the remaining functions func_B and func_C are irrelevant to the user query. This assumption is not validated in the paper. If the remaining functions are semantically overlapping or generic enough to answer the query, the generated 'answer directly' labels are incorrect, and the measured Relevance Detection gains from NF-1k could reflect label noise rather than improved relevance detection. Please provide a validation procedure, such as human or model-based checks of the NF-1k labels, and report the estimated label accuracy.
  4. [4.1 and 4.5, Tables 1-3] All conclusions are based on point estimates from a single base model (Breeze-7B) without confidence intervals, significance tests, or multiple random seeds, yet the prose uses language such as 'significantly enhances' and 'significant improvements.' This is especially concerning for the translation claim in Table 3, where the ZHTW benchmark is constructed by Lee et al. (2024), whose author list includes two authors of this paper. Please report variance information (e.g., seeds or bootstrap CIs), temper the significance language accordingly, and disclose or mitigate the potential benchmark-author overlap by using an independent evaluation where possible.
minor comments (4)
  1. [4.4] There is a typo in 'trained onIF-110k + FC-110k-Reason'; a space is missing after 'on'.
  2. [4.5] The phrase 'an non-function-call case generation pipeline' contains a grammatical error; it should be 'a non-function-call case generation pipeline.'
  3. [Table 1] The 'Use of Data?' columns use check and cross symbols, but the symbols are not defined in the caption or the text. Please define them explicitly.
  4. [4.1] The MT-Bench scores in Table 1 are reported as point estimates from a GPT-4o judge, but no information is given about judge stability or variance. A brief note or additional runs would improve confidence in the instruction-following comparison.

Circularity Check

1 steps flagged · score 4.0 of 10

No circular reduction in the English claims; multilingual evaluation rests on the authors' own ZHTW benchmark, raising a moderate self-citation concern.

  1. self citation load bearing [Section 4.1, Experimental Setup; Table 3]
    "We also evaluated the performance on Traditional Chinese function calling using the Function Calling Leaderboard for ZHTW (Lee et al., 2024), which is constructed by translating the BFCL. Therefore, the calculation of metrics AST Summary and Relevance Detection is similar."

    The multilingual contribution (claim 3) is evaluated exclusively on the ZHTW benchmark of Lee et al. (2024), whose author list includes Yi-Chang Chen and Da-shan Shiu, two of the present paper's authors. All Traditional Chinese numbers in Table 3 come from this benchmark, and the paper provides no external validation of it, so the translation-pipeline conclusion rests on a self-citation. This is not a definitional circularity because the benchmark is a translation of the external BFCL, but it is load-bearing for a secondary claim and therefore raises the score moderately.

full rationale

The central claims are empirical comparisons on external benchmarks: Table 1 uses BFCL (Yan et al., 2024) and MT-Bench (Zheng et al., 2023), and Table 2 uses BFCL relevance detection. The instruction-following comparison (rows b vs d) is quantity-confounded because adding IF-110k also doubles the training corpus, but that is a validity threat rather than a circular reduction: the BFCL metrics are not defined in terms of the training data, and the improvement could in principle fail. The Decision Token + NF-1k result is likewise tested on the external BFCL; although NF-1k is constructed from the same 'irrelevant functions' assumption, the test is not defined in terms of the token, so no by-construction equivalence exists. No fitted parameter is renamed as a prediction, and no load-bearing uniqueness theorem or ansatz is imported from prior work. The only notable self-citation is the ZHTW benchmark (Lee et al., 2024), built by two of the present authors, which supports the secondary multilingual claim; because it is a translation of the external BFCL, it retains some independent grounding, so the issue is moderate rather than disqualifying.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

All findings are empirical, not derivational, so the ledger records manually chosen data volumes that condition the claims, the domain assumptions about benchmarks and labels, and the Decision Token as the one new mechanism introduced. No numerically fitted statistical parameters appear in the paper.

free parameters (4)
  • IF-110k volume = 110,000
    Instruction-following examples sampled from Open ORCA; the claimed benefit of instruction-following data is established only at this volume, not across a range.
  • FC-110k volume = 110,000
    Function-calling examples from APIGen and glaive-function-calling-v2; held constant across all experiments.
  • NF-1k volume = 1,000
    Synthetic non-function-call examples used with the Decision Token; the relevance detection gain is reported only at this volume.
  • TC-19k volume = 19,000
    Traditional Chinese translated function-calling examples; the multilingual improvement is reported at this single volume.
assumptions (5)
  • domain assumption Breeze-7B is a representative base model for evaluating function-calling enhancements.
    All experiments use a single 7B model from the authors' own lab (Hsu et al., 2024); the paper does not test other base models, so generalization is assumed.
  • domain assumption BFCL AST Summary and Relevance Detection are valid and sufficient measures of function-calling ability.
    Scores are taken from the Berkeley Function Calling Leaderboard as defined in Yan et al., 2024; the paper does not analyze failure cases or additional metrics.
  • domain assumption Synthetic non-function-call labels are correct when the helpful function is removed.
    Section 3.2 assumes remaining functions (func_B, func_C) are not helpful after removing func_A; if overlapping or generic functions are present, the 'answer directly' label is wrong.
  • domain assumption MT-Bench scores from a GPT-4o judge reliably reflect instruction-following quality.
    The paper uses GPT-4o as judge following Zheng et al., 2023; judge bias is not assessed.
  • domain assumption The ZHTW benchmark (Lee et al., 2024) is a faithful Traditional Chinese translation of BFCL.
    Multilingual results are evaluated on a benchmark built by two of this paper's authors; its validity is assumed rather than independently verified.
invented entities (1)
  • Decision Token (<|answer|>, <|use_tool|>)
    purpose: Binary decision marker before response generation; enables low-cost synthetic non-function-call labels for relevance detection.
    The token is proposed and tested only in this paper; no code, checkpoints, or external replication are provided, so its effectiveness rests on the paper's own reported BFCL numbers.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Function-Calling Capabilities in LLMs: Strategies for Prompt Formats, Data Integration, and Multilingual Translation." pith.science (2026). https://pith.science/paper/WS7E3VKD

@misc{pith2026241201130,
  author       = {Pith},
  title        = {Pith review of: Enhancing Function-Calling Capabilities in LLMs: Strategies for Prompt Formats, Data Integration, and Multilingual Translation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WS7E3VKD}},
  note         = {Machine review of arXiv:2412.01130}
}
read the original abstract

Large language models (LLMs) have significantly advanced autonomous agents, particularly in zero-shot tool usage, also known as function calling. This research delves into enhancing the function-calling capabilities of LLMs by exploring different approaches, including prompt formats for integrating function descriptions, blending function-calling and instruction-following data, introducing a novel Decision Token for conditional prompts, leveraging chain-of-thought reasoning, and overcoming multilingual challenges with a translation pipeline. Our key findings and contributions are as follows: (1) Instruction-following data improves both function-calling accuracy and relevance detection. (2) The use of the newly proposed Decision Token, combined with synthetic non-function-call data, enhances relevance detection. (3) A tailored translation pipeline effectively overcomes multilingual limitations, demonstrating significant improvements in Traditional Chinese. These insights highlight the potential for improved function-calling capabilities and multilingual applications in LLMs.

Figures

Figures reproduced from arXiv: 2412.01130 by the authors.

Figure 1
Figure 1. An illustration of prompt templates used for function calling and instruction following in LLMs. Dur [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. RAG-MCP: Mitigating Prompt Bloat in LLM Tool Selection via Retrieval-Augmented Generation

    cs.AI 2025-05 reject novelty 4.0 of 10

    RAG-MCP retrieves the most relevant Model Context Protocol server description for each query and feeds only that schema to the LLM, reporting fewer prompt tokens and higher selection accuracy on a small benchmark.

  2. The Breeze 2 Herd of Models: Traditional Chinese LLMs Based on Llama with Vision-Aware and Function-Calling Capabilities

    cs.CL 2025-01 conditional novelty 3.0 of 10

    Breeze2 continues pretraining Llama on a 900GB Traditional Chinese corpus, then adds LLaVA-style vision and function-calling fine-tuning, reporting top size-class scores on a mix of public and self-built benchmarks.

Reference graph

Works this paper leans on

35 extracted references · 8 canonical work pages · cited by 2 Pith papers

  1. [1]

    Ibrahim Abdelaziz, Kinjal Basu, Mayank Agarwal, Sadhana Kumaravel, Matthew Stallone, Rameswar Panda, Yara Rizk, GP Bhargav, Maxwell Crouse, Chulaka Gunasekara, Shajith Ikbal, Sachin Joshi, Hima Karanam, Vineet Kumar, Asim Munawar, Sumit Neelam, Dinesh Raghu, Udit Sharma, Adriana Meza Soria, Dheeraj Sreedhar, Praveen Venkateswaran, Merve Unuvar, David Cox,...

    arXiv 2024

  2. [2]

    Maxwell Crouse, Ibrahim Abdelaziz, Kinjal Basu, Soham Dan, Sadhana Kumaravel, Achille Fokoue, Pavan Kapanipathi, and Luis A. Lastras. 2024. https://openreview.net/forum?id=FRxDrdysBt Formally specifying the high-level behavior of LLM -based agents

    work page 2024

  3. [3]

    Luyu Gao, Aman Madaan, Shuyan Zhou, Uri Alon, Pengfei Liu, Yiming Yang, Jamie Callan, and Graham Neubig. 2022. Pal: Program-aided language models. arXiv preprint arXiv:2211.10435

    arXiv 2022

  4. [4]

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, Amy Yang, Angela Fan, Anirudh Goyal, Anthony Hartshorn, Aobo Yang, Archi Mitra, Archie Sravankumar, Artem Korenev, Arthur Hinsvark, Arun Rao, Aston Zhang, Aurelien Rodriguez, Austen Gregerson, Ava S...

    arXiv 2024

  5. [5]

    Izzeddin Gur, Hiroki Furuta, Austin V Huang, Mustafa Safdari, Yutaka Matsuo, Douglas Eck, and Aleksandra Faust. 2024. https://openreview.net/forum?id=9JQtrumvg8 A real-world webagent with planning, long context understanding, and program synthesis . In The Twelfth International Conference on Learning Representations

    2024

  6. [6]

    Yilun Hao, Yongchao Chen, Yang Zhang, and Chuchu Fan. 2024. Large language models can plan your travels rigorously with formal verification tools. arXiv preprint arXiv:2404.11891

    arXiv 2024

  7. [7]

    Wang, and Noah D

    Joy He-Yueya, Gabriel Poesia, Rose E. Wang, and Noah D. Goodman. 2023. https://arxiv.org/abs/2304.09102 Solving math word problems by combining language models with symbolic solvers . Preprint, arXiv:2304.09102

    arXiv 2023

  8. [8]

    Chan-Jan Hsu, Chang-Le Liu, Feng-Ting Liao, Po-Chun Hsu, Yi-Chang Chen, and Da-Shan Shiu. 2024. Breeze-7b technical report. arXiv preprint arXiv:2403.02712

    arXiv 2024

Show all 35 references
  1. [9]

    Shijue Huang, Wanjun Zhong, Jianqiao Lu, Qi Zhu, Jiahui Gao, Weiwen Liu, Yutai Hou, Xingshan Zeng, Yasheng Wang, Lifeng Shang, Xin Jiang, Ruifeng Xu, and Qun Liu. 2024. https://doi.org/10.18653/v1/2024.findings-acl.259 Planning, creation, usage: Benchmarking LLM s for comprehe...

    2024 doi

  2. [10]

    Mojtaba Komeili, Kurt Shuster, and Jason Weston. 2021. https://api.semanticscholar.org/CorpusID:236034557 Internet-augmented dialogue generation . In Annual Meeting of the Association for Computational Linguistics

    2021

  3. [11]

    Liang-Chieh Lee, Cheng-Wei Lin, Pei-Chen Ho, Chien-Yu Yu, Yi-Chang Chen, and Da-Shan Shiu. 2024. https://github.com/mtkresearch/function-calling-leaderboard-for-zhtw Function calling leaderboard for zhtw

    2024

  4. [12]

    Weiwen Liu, Xu Huang, Xingshan Zeng, Xinlong Hao, Shuai Yu, Dexun Li, Shuai Wang, Weinan Gan, Zhengying Liu, Yuanqing Yu, Zezhong Wang, Yuxian Wang, Wu Ning, Yutai Hou, Bin Wang, Chuhan Wu, Xinzhi Wang, Yong Liu, Yasheng Wang, Duyu Tang, Dandan Tu, Lifeng Shang, Xin Jiang, Rui...

    2024 arXiv

  5. [13]

    Zuxin Liu, Thai Hoang, Jianguo Zhang, Ming Zhu, Tian Lan, Shirley Kokane, Juntao Tan, Weiran Yao, Zhiwei Liu, Yihao Feng, et al. 2024 b . Apigen: Automated pipeline for generating verifiable and diverse function-calling datasets. arXiv preprint arXiv:2406.18518

    2024 arXiv

  6. [14]

    Shayne Longpre, Le Hou, Tu Vu, Albert Webson, Hyung Won Chung, Yi Tay, Denny Zhou, Quoc V Le, Barret Zoph, Jason Wei, et al. 2023. The flan collection: Designing data and methods for effective instruction tuning. In International Conference on Machine Learning, pages 22631--22...

    2023

  7. [15]

    Nexusflow.ai. 2023. https://nexusflow.ai/blogs/ravenv2 Nexusraven-v2: Surpassing gpt-4 for zero-shot function calling

    2023

  8. [16]

    Aaron Parisi, Yao Zhao, and Noah Fiedel. 2022. https://arxiv.org/abs/2205.12255 Talm: Tool augmented language models . Preprint, arXiv:2205.12255

    2022 arXiv

  9. [17]

    Patil, Tianjun Zhang, Xin Wang, and Joseph E

    Shishir G. Patil, Tianjun Zhang, Xin Wang, and Joseph E. Gonzalez. 2023. Gorilla: Large language model connected with massive apis. arXiv preprint arXiv:2305.15334

    2023 arXiv

  10. [18]

    Yujia Qin, Shihao Liang, Yining Ye, Kunlun Zhu, Lan Yan, Yaxi Lu, Yankai Lin, Xin Cong, Xiangru Tang, Bill Qian, Sihan Zhao, Lauren Hong, Runchu Tian, Ruobing Xie, Jie Zhou, Mark Gerstein, dahai li, Zhiyuan Liu, and Maosong Sun. 2024. https://openreview.net/forum?id=dHng2O0Jjr...

    2024

  11. [19]

    Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xua, and Ji-Rong Wen. 2024. Tool learning with large language models: A survey. arXiv preprint arXiv:2405.17935

    2024 arXiv

  12. [20]

    Timo Schick, Jane Dwivedi-Yu, Roberto Dessi, Roberta Raileanu, Maria Lomeli, Eric Hambro, Luke Zettlemoyer, Nicola Cancedda, and Thomas Scialom. 2023. https://openreview.net/forum?id=Yacmpz84TH Toolformer: Language models can teach themselves to use tools . In Thirty-seventh C...

    2023

  13. [21]

    Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. 2023. https://arxiv.org/abs/2303.11366 Reflexion: Language agents with verbal reinforcement learning . Preprint, arXiv:2303.11366

    2023 arXiv

  14. [22]

    Qiaoyu Tang, Ziliang Deng, Hongyu Lin, Xianpei Han, Qiao Liang, and Le Sun. 2023. https://arxiv.org/abs/2306.05301 Toolalpaca: Generalized tool learning for language models with 3000 simulated cases . Preprint, arXiv:2306.05301

    2023 arXiv

  15. [23]

    Ryan Teknium, Jeffrey Quesnelle, and Chen Guang. 2024. https://arxiv.org/abs/2408.11857 Hermes 3 technical report . Preprint, arXiv:2408.11857

    2024 arXiv

  16. [24]

    Adrian Theuma and Ehsan Shareghi. 2024. https://arxiv.org/abs/2401.15328 Equipping language models with tool use capability for tabular data analysis in finance . Preprint, arXiv:2401.15328

    2024 arXiv

  17. [25]

    Boshi Wang, Hao Fang, Jason Eisner, Benjamin Van Durme, and Yu Su. 2024. https://doi.org/10.18653/v1/2024.acl-long.570 LLM s in the imaginarium: Tool learning through simulated trial and error . In Proceedings of the 62nd Annual Meeting of the Association for Computational Lin...

    2024 doi

  18. [26]

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

    2022

  19. [27]

    Binfeng Xu, Zhiyuan Peng, Bowen Lei, Subhabrata Mukherjee, Yuchen Liu, and Dongkuan Xu. 2023. https://arxiv.org/abs/2305.18323 Rewoo: Decoupling reasoning from observations for efficient augmented language models . Preprint, arXiv:2305.18323

    2023 arXiv

  20. [28]

    Patil, Ion Stoica, and Joseph E

    Fanjia Yan, Huanzhi Mao, Charlie Cheng-Jie Ji, Tianjun Zhang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez. 2024. Berkeley function calling leaderboard. https://gorilla.cs.berkeley.edu/blogs/8_berkeley_function_calling_leaderboard.html

    2024

  21. [29]

    Rui Yang, Lin Song, Yanwei Li, Sijie Zhao, Yixiao Ge, Xiu Li, and Ying Shan. 2023 a . https://openreview.net/forum?id=cwjh8lqmOL GPT 4tools: Teaching large language model to use tools via self-instruction . In Thirty-seventh Conference on Neural Information Processing Systems

    2023

  22. [30]

    Zhengyuan Yang, Linjie Li, Jianfeng Wang, Kevin Lin, Ehsan Azarnasab, Faisal Ahmed, Zicheng Liu, Ce Liu, Michael Zeng, and Lijuan Wang. 2023 b . Mm-react: Prompting chatgpt for multimodal reasoning and action

    2023

  23. [31]

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2022. React: Synergizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629

    2022 arXiv

  24. [32]

    Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena. Advances in Neural Information Processing Systems, 36:46595--46623

    2023

  25. [33]

    Ruizhe Zhong, Xingbo Du, Shixiong Kai, Zhentao Tang, Siyuan Xu, Hui-Ling Zhen, Jianye Hao, Qiang Xu, Mingxuan Yuan, and Junchi Yan. 2023. https://arxiv.org/abs/2401.12224 Llm4eda: Emerging progress in large language models for electronic design automation . Preprint, arXiv:2401.12224

    2023 arXiv

  26. [34]

    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...

  27. [35]

    write newline

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

Pith tools

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