Papers › From Exploration to Mastery: Enabling LLMs to Master Tools via Self-Driven Interactions

From Exploration to Mastery: Enabling LLMs to Master Tools via Self-Driven Interactions

10 Oct 2024arXiv:2410.08197archive 2025-07-28

Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, Ji-Rong Wen

Tool learning enables Large Language Models (LLMs) to interact with external environments by invoking tools, serving as an effective strategy to mitigate the limitations inherent in their pre-training data. In this process, tool documentation plays a crucial role by providing usage instructions for LLMs, thereby facilitating effective tool utilization. This paper concentrates on the critical challenge of bridging the comprehension gap between LLMs and external tools due to the inadequacies and inaccuracies inherent in existing human-centric tool documentation. We propose a novel framework, DRAFT, aimed at Dynamically Refining tool documentation through the Analysis of Feedback and Trails emanating from LLMs' interactions with external tools. This methodology pivots on an innovative trial-and-error approach, consisting of three distinct learning phases: experience gathering, learning from experience, and documentation rewriting, to iteratively enhance the tool documentation. This process is further optimized by implementing a diversity-promoting exploration strategy to ensure explorative diversity and a tool-adaptive termination mechanism to prevent overfitting while enhancing efficiency. Extensive experiments on multiple datasets demonstrate that DRAFT's iterative, feedback-based refinement significantly ameliorates documentation quality, fostering a deeper comprehension and more effective utilization of tools by LLMs. Notably, our analysis reveals that the tool documentation refined via our approach demonstrates robust cross-model generalization capabilities.

PaperPDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Syntology holds a reference implementation of DRAFT, generated from the paper text (not the authors' code), at level 3 on Syntology's ladder: V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” (quoted from Syntology's server card). No rung compares this code against the authors' implementation. For agents: get_reference_implementation(name='DRAFT') (how to connect); the free check have('2410.08197') needs no token.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 27 code samples harvested from 1 repository linked to this paper: 17 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 2 honoured, 1 violated, 1 with no contract checked; 13 where Syntology's instrument failed) · 10 unverified. Where Syntology's instrument failed: 13 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 27 samples from 1 repository, 17 ran (4 with no instrument failure; 13 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 7 of the 17 samples that ran have no recorded date.

quchangle1/DRAFT officialmentioned on GitHubpytorchSyntology: 27 samples, 17 ran (13 instrument-failed) · 10 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

27 samples harvested: 17 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 2 honoured, 1 violated, 1 with no contract checked; 13 where Syntology's instrument failed) · 10 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran · honoured contract
1ran · violated contract
13ran · our draft was wrong
1ran
10unverified

Licence: 0 of the 27 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from quchangle1/DRAFT. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

quchangle1/DRAFT official repository 27 samples, 17 ran (13 instrument-failed) · 10 unverified
Info Inference_DFSDT.py ran MIT (permissive) · 90652e52f646264e · report fixture: returned dict · level recorded 2026-09-24
answer_check Inference_DFSDT.py ran · honoured contract MIT (permissive) · 8fa191b03e06ca45 · report fixture: returned int · level recorded 2026-09-24
change_name Cal_path_rate.py ran · honoured contract fingerprinted MIT (permissive) · 80196c9b6d0a1371 · report fixture: signature name:str/0/str · returned str · level recorded, date not recorded
change_name Inference_DFSDT.py ran · our draft was wrong fingerprinted MIT (permissive) · b273ba46971479bb · report fixture: signature name:str/0/str · returned str · level recorded 2026-09-22
choose_parameter Inference_DFSDT.py ran · our draft was wrong fingerprinted MIT (permissive) · 20b0113771cb17f2 · report fixture: signature API_instruction:str/0/str, question:str/0/str, model_name:str/0/str · returned dict · level recorded 2026-09-24
choose_parameter_depend Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · 6925fcd2a5d56794 · report fixture: signature API_instruction:str/0/str, question:str/0/str, previous_log:str/0/str, model_name:str/0/str · returned dict · level recorded 2026-09-24
choose_tool Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · e0e1e1a4f708f8c7 · report fixture: signature question:str/0/str, Tool_dic:list/1/dict, tool_used:set/0/?, model_name:str/0/str · returned dict · level recorded 2026-09-24
cosine_similarity in 120 papers
census: 26 implementations from 30 papers, 11 distinct outputs across 5 buckets, one shared input per bucket
DRAFT.py ran · violated contract fingerprinted MIT (permissive) · fe3dcc49d4415434 · report fixture: signature vec1:list/1/float, vec2:list/1/float · returned float64 · level recorded, date not recorded
dict_shorten Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · 1baad03523efe558 · report fixture: signature origin:dict/0/?, schema:dict/0/? · returned dict · level recorded 2026-09-24
get_response in 48 papers Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · e247a2919937c8b6 · report fixture: signature messages:list/1/dict, temperature:float/0/float, top_p:int/0/int, max_tokens:int/0/int, model:str/0/str · returned str · level recorded 2026-09-24
observation_shorten Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · a3869e242f4b29e5 · report fixture: signature schema_root:str/0/str, response_dict:dict/0/?, category:str/0/str, tool_name:str/0/str, api_name:str/0/str, strip_method:str/0/str · returned str · level recorded 2026-09-24
prepare_tool_name_and_url Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · 99eef45bc570d9d0 · report fixture: signature tools_root:str/0/str, info:MockInfo/0/? · returned tuple · level recorded, date not recorded
process_error Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · 2e8458283ddaa9b4 · report fixture: signature response:MockResponse/0/? · returned tuple · level recorded, date not recorded
process_name Cal_path_rate.py ran · our draft was wrong fingerprinted MIT (permissive) · 9103b4347a03d6f5 · report fixture: signature name:str/0/str · returned str · level recorded, date not recorded
run in 143 papers Inference_DFSDT.py ran · our draft was wrong fingerprinted MIT (permissive) · 8a54f4acce793e65 · report fixture: signature toolbench_code_string:str/0/str, toolbench_api_name:str/0/str, toolbench_input_params_str:str/0/str · returned tuple · level recorded, date not recorded
standardize in 47 papers Cal_path_rate.py ran · our draft was wrong fingerprinted MIT (permissive) · c25f57d20852531f · report fixture: signature string:str/0/str · returned str · level recorded, date not recorded
task_decompose Inference_DFSDT.py ran · our draft was wrong MIT (permissive) · 1f9fe0a7c687f7f6 · report fixture: signature question:str/0/str, Tool_dic:dict/0/?, model_name:str/0/str · returned dict · level recorded 2026-09-24
answer_generation Inference_DFSDT.py unverified MIT (permissive) · 4a851234c230b4f3 · report
answer_generation_depend Inference_DFSDT.py unverified MIT (permissive) · 166d885b223fec87 · report
answer_summarize Inference_DFSDT.py unverified MIT (permissive) · f092dbfafd36bcc8 · report
choose_API Inference_DFSDT.py unverified MIT (permissive) · e1ed82e16201c68a · report
get_rapidapi_response Inference_DFSDT.py unverified MIT (permissive) · 54fedb6237f6f7d5 · report
openai_response Inference_DFSDT.py unverified MIT (permissive) · 721a78b42e639559 · report
retrieval Inference_DFSDT.py unverified MIT (permissive) · cdac1abe959ba09c · report
task_execution Inference_DFSDT.py unverified MIT (permissive) · 8be639625e74f9fc · report
task_topology Inference_DFSDT.py unverified MIT (permissive) · 19a64bf2fd72679c · report
update_progress Inference_DFSDT.py unverified MIT (permissive) · 479a78407b99eb56 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. For a generated assembly, an independent peer is a second implementation generated separately from the same paper text; neither implementation read the authors' code. At level 2 the two agreed numerically on the fixture's inputs; level 3 is recorded when property tests derived from the paper also passed on both, and each level 3 card prints how many its report lists. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. A generated assembly's label was set when its code was generated for the method, and no model judged it; its card prints it as "by construction" or "not judged by a model". Where an assembly's phrase has no constructed figure, a brick that ran is not among the harvested samples this build reads, and whether a brick's run only constructed an object is read from that brick's harvested sample, so the figure was not counted; it does not say there were none.

2 assemblies on this paper: 1 harvested, 1 generated.

generated Assembly for DRAFT (graph method: DRAFT), generated from the paper text, not the repository's code. 1 brick; individually on a fixture: 1 ran (0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified. Ran end-to-end against an independent peer (level 3; 6 property tests passed on both).

label matches by construction: the code was generated for this method (not judged by a model)

  1. calculate_documentation_similarity entry · level 3 · ran · our draft was wrong · 06c4fd1d3017514cd88ada79f19d64aa829a564ebd9f2a04ac386e91598a37bd

harvested Assembly for DRAFT (graph method: DRAFT). 24 bricks; individually on a fixture: 14 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 12 where Syntology's instrument failed) · 10 unverified. The assembly has not been run end-to-end.

  1. task_execution entry · level 0 · unverified · 8be639625e74f9fc
  2. answer_check metric · level 1 · ran · honoured contract · 8fa191b03e06ca45
  3. answer_generation helper · level 0 · unverified · 4a851234c230b4f3
  4. answer_generation_depend helper · level 0 · unverified · 166d885b223fec87
  5. answer_summarize helper · level 0 · unverified · f092dbfafd36bcc8
  6. retrieval helper · level 0 · unverified · cdac1abe959ba09c
  7. task_decompose helper · level 1 · ran · our draft was wrong · 1f9fe0a7c687f7f6
  8. task_topology helper · level 0 · unverified · 19a64bf2fd72679c
  9. update_progress helper · level 0 · unverified · 479a78407b99eb56
  10. openai_response helper · level 0 · unverified · 721a78b42e639559
  11. change_name helper · level 1 · ran · our draft was wrong · b273ba46971479bb
  12. choose_API helper · level 0 · unverified · e1ed82e16201c68a
  13. choose_parameter helper · level 1 · ran · our draft was wrong · 20b0113771cb17f2
  14. choose_parameter_depend helper · level 1 · ran · our draft was wrong · 6925fcd2a5d56794
  15. choose_tool helper · level 1 · ran · our draft was wrong · e0e1e1a4f708f8c7
  16. get_rapidapi_response helper · level 0 · unverified · 54fedb6237f6f7d5
  17. standardize helper · level 1 · ran · our draft was wrong · e6d1b09cb3624d3f
  18. get_response helper · level 1 · ran · our draft was wrong · e247a2919937c8b6
  19. Info helper · level 1 · ran · 90652e52f646264e
  20. observation_shorten helper · level 1 · ran · our draft was wrong · a3869e242f4b29e5
  21. prepare_tool_name_and_url helper · level 1 · ran · our draft was wrong · 99eef45bc570d9d0
  22. run helper · level 1 · ran · our draft was wrong · 8a54f4acce793e65
  23. dict_shorten helper · level 1 · ran · our draft was wrong · 1baad03523efe558
  24. process_error helper · level 1 · ran · our draft was wrong · 2e8458283ddaa9b4

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2410.08197") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Diversity

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections