SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.14751 · NeurIPS · 2024

AGILE: A Novel Reinforcement Learning Framework of LLM Agents

Hang Li, Peiyuan Feng, Yichen He, Guanhua Huang, Yuan Lin, Hanchong Zhang, Yuchen Zhang

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 4 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2405.14751.json · badge: /api/ran/2405.14751.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
bytarnish/AGILE canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
FunctionStatusWhere it lives
chunks Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
bytarnish/AGILE/agile/data_processing/fact_qa_gen.py
code served (permissive licence) · get_code("29ea8bc80be13db5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Analyzing the function:
# - `chunks` takes a list `lst` and an integer `n`
# - It splits `lst` into chunks of size `n`
# - It returns a list of sublists
# - The function is simple and only needs a list and an integer

lst = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
n = 3

ARGS = {
    "lst": lst,
    "n": n
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_action Ran
ran; our contract draft was wrong, not the code
bytarnish/AGILE/agile/hotpot_agent.py
code served (permissive licence) · get_code("4f65da0e3cfe9f59")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

answer_pattern = 'Action: \\[(.*?)\\] \\((.*?)\\)'

def get_action(res):
    matches = re.findall(answer_pattern, res, re.DOTALL)
    if len(matches) < 1:
        return (None, '', '')
    prefix = res.split('[{}] ({})'.format(matches[0][0], matches[0][1]))[0]
    if matches[0][0].lower() == 'search':
        return ('Search', matches[0][1], prefix)
    elif matches[0][0].lower() == 'predictanswer':
        return ('PredictAnswer', matches[0][1], prefix)
    else:
        return ('SeekAdvice', '', prefix)


# Example 1: Valid Search action
res_search = "Some prefix text Action: [Search] (query_term)"

# Example 2: Valid PredictAnswer action
res_predict = "Initial context Action: [PredictAnswer] (the final answer)"

# Example 3: Valid SeekAdvice action
res_advice = "Preamble Action: [UnknownAction] (something)"

# Example 4: No match case
res_no_match = "This has no action pattern"

# Test with one of them
test_res = res_search

ARGS = {
    'res': test_res
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_group_split Ran
ran; our contract draft was wrong, not the code
bytarnish/AGILE/agile/utils.py
code served (permissive licence) · get_code("81ab3ea5730888f3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# The function get_group_split takes a single parameter 'group'
# It checks if group is in a list of strings and returns 'test' or 'train'
# So group should be a string

ARGS = {
    'group': 'all_pans'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mean_pooling Ran
ran; our contract draft was wrong, not the code
bytarnish/AGILE/agile/models.py
pointer only (licence: MIT); the stored copy get_code serves, from dengyang17/llm-proactive, is pointer only (licence: NONE) · get_code("77f69af078e09830")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from dengyang17/llm-proactive, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 1 later) and 3 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
mean_pooling 22 other papers
15 dated earlier, 1 the same month, 6 later · 1 only as an assembly brick
2206.04624 (2022-06)
Factuality Enhanced Language Models for Open-Ended Text Generation
2023.findings-emnlp.540 (2023 (year only))
Open-ended Commonsense Reasoning with Unrestricted Answer Scope
2302.05698 (2023-02)
Compositional Exemplars for In-context Learning
and 19 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We introduce a novel reinforcement learning framework of LLM agents named AGILE (AGent that Interacts and Learns from Environments) designed to perform complex conversational tasks with users, leveraging LLMs, memory, tools, and interactions with experts. The agent possesses capabilities beyond conversation, including reflection, tool usage, and expert consultation. We formulate the construction of such an LLM agent as a reinforcement learning (RL) problem, in which the LLM serves as the policy model. We fine-tune the LLM using labeled data of actions and the PPO algorithm. We focus on question answering and release a dataset for agents called ProductQA, comprising challenging questions in online shopping. Our extensive experiments on ProductQA, MedMCQA and HotPotQA show that AGILE agents based on 7B and 13B LLMs trained with PPO can outperform GPT-4 agents. Our ablation study highlights the indispensability of memory, tools, consultation, reflection, and reinforcement learning in achieving the agent's strong performance. Datasets and code are available at https://github.com/bytarnish/AGILE.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2405.14751")
get_code_for_paper("2405.14751")
have("2405.14751")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2405.14751.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2405.14751.svg)](https://syntology.ai/paper/2405.14751)

Connect an agent — have() is free.