SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2401.03428 · 2024

Exploring Large Language Model based Intelligent Agents: Definitions, Methods, and Prospects

arXiv · PDF · Open in the Atlas

Code that ran

7 of 8 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, 1 violated it and 0 have no contract verdict; 6 failures of our instrument; 1 no recorded run), measured 2026-11-21; 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/2401.03428.json · badge: /api/ran/2401.03428.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
melih-unsal/demogpt canonical (classifier label)
classifier basis: census
7 ran · 1 no recorded run
FunctionStatusWhere it lives
get_prompt Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/prompt.py
code served (permissive licence) · get_code("938213d47dbb3323")
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 ChatPromptTemplate. 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
from langchain_core.prompts.chat import ChatPromptTemplate, HumanMessagePromptTemplate, SystemMessagePromptTemplate

torch.manual_seed(0)

# The function takes two string parameters: system_template and human_template
# Both are used with .from_template() method which expects string templates
# The function checks if they are truthy before using them

system_template = "You are a helpful assistant. Answer the user's question in {language}."
human_template = "User question: {question}"

ARGS = {
    "system_template": system_template,
    "human_template": human_template,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
langchain_core, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ChatPromptTemplate — a ChatPromptTemplate 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
inputs_joiner Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/utils.py
code served (permissive licence) · get_code("5f83041fb05b1b22")
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

def inputs_joiner(var):
    """
    used in generating if statement. Instead of if var1 and var2 and ..., it also checks if it is bool.
    """
    return f'(isinstance({var},bool) or {var})'

# The function takes a single parameter 'var' which is a string (variable name)
# The function uses string formatting with f-strings to construct a conditional check
# No actual runtime behavior depends on the value of var - it's just used as a string in formatting
# So var should be a string representing a variable name

ARGS = {
    "var": "my_variable"
}
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
isTaskAvailable Ran
ran, violated the contract we drafted
melih-unsal/demogpt/demogpt/chains/task_definitions.py
code served (permissive licence) · get_code("62932bcb4ba99e37")
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 bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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

# Based on the function, we need to construct:
# - task: a dict with a 'name' key
# - app_chat: a boolean
# - app_prompt_template: a boolean
# - app_search: a boolean
# - app_summary: a boolean

# The function only accesses task['name'] as a string and checks boolean flags
# Let's create valid test arguments

task = {
    'name': 'chat'
}

app_chat = True
app_prompt_template = True
app_search = True
app_summary = True

ARGS = {
    'task': task,
    'app_chat': app_chat,
    'app_prompt_template': app_prompt_template,
    'app_search': app_search,
    'app_summary': app_summary
}
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 bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
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
jsonFixer Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/chains/task_definitions.py
code served (permissive licence) · get_code("b0089960553ecb91")
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 json

def jsonFixer(data):
    data = json.dumps(data, indent=4)
    return data.replace('{', '{{').replace('}', '}}')

ARGS = {
    'data': {
        'name': 'example',
        'values': [1, 2, 3],
        'nested': {'key': 'value'}
    }
}
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 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
planToTaskFormat Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/controllers.py
code served (permissive licence) · get_code("987d4bf3882aaeb2")
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, 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

def planToTaskFormat(plan):
    pattern = re.compile('\\b(\\w+)\\(([^)]*)\\)( ---> ([\\w_,\\s]+))?')
    matches = pattern.findall(plan)
    functions = [{'name': match[0], 'arguments': [arg.strip() for arg in match[1].split(',') if arg.strip() and '"' not in arg.strip()], 'output': [arg.strip() for arg in match[3].split(',') if arg.strip() and '"' not in arg.strip()]} for match in matches]
    plan_steps = []
    for i, function in enumerate(functions):
        step = {'step': i + 1, 'task_type': function['name'], 'task_name': function['name'], 'input_key': function['arguments'], 'output_key': function['output']}
        plan_steps.append(step)
    return plan_steps

# Test with a valid plan string
plan_string = "retrieve(query, context) ---> results, confidence\nprocess(results) ---> output"

ARGS = {
    'plan': plan_string
}
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 list, ndim 1, dtype family float — a list 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
separateCode Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/utils.py
code served (permissive licence) · get_code("99828e1c6cde2e68")
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 ast

# Create a small Python source code string with imports, functions, and other statements
source = """
import os
from collections import defaultdict

def add(a, b):
    return a + b

def multiply(x, y):
    return x * y

x = 10
y = 20
z = x + y

if z > 25:
    result = z * 2
else:
    result = z

for i in range(5):
    print(i)

class MyClass:
    pass
"""

ARGS = {
    "source": source
}
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
separateCode Ran
ran; our contract draft was wrong, not the code
melih-unsal/demogpt/demogpt/utils.py
code served (permissive licence) · get_code("9686907319d548b9")
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 dict. 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 re

code_sample = """import os
import sys
from pathlib import Path

def hello(name):
    print(f"Hello {name}")
    return name.upper()

def add(x, y):
    return x + y

x = 5
y = 10
result = x + y
"""

ARGS = {
    'code': code_sample
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
getAvailableTasks Not yet run
unverified (harvested, no recorded run)
melih-unsal/demogpt/demogpt/chains/task_definitions.py
code served (permissive licence) · get_code("c571772246034707")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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.

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 2026-11-21.

Repositories linked to this paper

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

Abstract

Intelligent agents stand out as a potential path toward artificial general intelligence (AGI). Thus, researchers have dedicated significant effort to diverse implementations for them. Benefiting from recent progress in large language models (LLMs), LLM-based agents that use universal natural language as an interface exhibit robust generalization capabilities across various applications -- from serving as autonomous general-purpose task assistants to applications in coding, social, and economic domains, LLM-based agents offer extensive exploration opportunities. This paper surveys current research to provide an in-depth overview of LLM-based intelligent agents within single-agent and multi-agent systems. It covers their definitions, research frameworks, and foundational components such as their composition, cognitive and planning methods, tool utilization, and responses to environmental feedback. We also delve into the mechanisms of deploying LLM-based agents in multi-agent systems, including multi-role collaboration, message passing, and strategies to alleviate communication issues between agents. The discussions also shed light on popular datasets and application scenarios. We conclude by envisioning prospects for LLM-based agents, considering the evolving landscape of AI and natural language processing.

For agents

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

get_harvested_code_for_paper("2401.03428")
get_code_for_paper("2401.03428")
have("2401.03428")

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

curl https://syntology.ai/api/ran/2401.03428.json

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

[![Syntology run record](https://syntology.ai/api/ran/2401.03428.svg)](https://syntology.ai/paper/2401.03428)

Connect an agent — have() is free.