SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.15057 · ACL · 2024

On the Multi-turn Instruction Following for Conversational Web Agents

Xuan Zhang, Tat-Seng Chua, Wenxuan Zhang, See-Kiong Ng, Yang Deng, Yifei Yuan

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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; 7 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/2402.15057.json · badge: /api/ran/2402.15057.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
magicgh/self-map canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
FunctionStatusWhere it lives
clean_text Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/dom_utils.py
code served (permissive licence) · get_code("0e07e2733eab1746")
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 re

def clean_text(text):
    if text is None:
        return ''
    text = re.sub('\\s+', ' ', text)
    return text.strip()

ARGS = {
    'text': '  hello   world  \n  foo  bar  '
}
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
convert_dict_to_dataclass Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/action_prediction/utils.py
code served (permissive licence) · get_code("aa62aebf8290e3a6")
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 OuterConfig. 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 dataclasses
import torch

torch.manual_seed(0)

@dataclasses.dataclass
class InnerConfig:
    value: int
    name: str

@dataclasses.dataclass
class OuterConfig:
    inner: InnerConfig
    count: int

# Test data: nested dict that matches the dataclass structure
test_data = {
    "inner": {
        "value": 42,
        "name": "test"
    },
    "count": 10
}

ARGS = {
    "cls": OuterConfig,
    "data": test_data
}
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 OuterConfig — a OuterConfig output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OuterConfig 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
encode_system_message Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/action_prediction/dataloader.py
code served (permissive licence) · get_code("3d5b29d6774c146f")
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 torch
import torch.nn as nn

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, text, add_special_tokens=True, truncation=True, max_length=None):
        # Simulate a tokenizer that returns a dict with 'input_ids' and 'attention_mask'
        tokens = [101] + [1000 + i for i in range(min(len(text.split()), max_length - 2 if max_length else 10))] + [102]
        if max_length:
            tokens = tokens[:max_length]
        return {
            'input_ids': tokens,
            'attention_mask': [1] * len(tokens)
        }

system_message = "You are a helpful assistant."
tokenizer = MockTokenizer()
max_context_len = 512

ARGS = {
    'system_message': system_message,
    'tokenizer': tokenizer,
    'max_context_len': max_context_len
}
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 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
fine_query Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/action_prediction/memory.py
code served (permissive licence) · get_code("4224dc92e5e82a52")
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)

# Construct valid arguments for fine_query function
# The function takes:
# - task: str
# - actions: Union[List[str], str]

ARGS = {
    "task": "classification",
    "actions": ["preprocess", "train", "evaluate"]
}
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
get_descendants Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/dom_utils.py
code served (permissive licence) · get_code("56fc0416b2cf42ae")
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.

class Node:
    def __init__(self, children=None):
        self.children = children if children is not None else []
    
    def __iter__(self):
        return iter(self.children)


# Build a small tree structure for testing
leaf1 = Node([])
leaf2 = Node([])
leaf3 = Node([])
middle1 = Node([leaf1, leaf2])
middle2 = Node([leaf3])
root = Node([middle1, middle2])

ARGS = {
    "node": root,
    "max_depth": 2,
    "current_depth": 0
}
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
load_pickle Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/action_prediction/utils.py
code served (permissive licence) · get_code("ea2042ec9678e7c1")
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 pickle
import tempfile
import os

# Create a temporary pickle file with some sample data
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "test_data.pkl")

sample_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "dict"}
}

with open(temp_file, 'wb') as f:
    pickle.dump(sample_data, f)

ARGS = {
    "file_name": temp_file
}
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
unsqueeze_turns Ran
ran; our contract draft was wrong, not the code
magicgh/self-map/src/action_prediction/memory.py
code served (permissive licence) · get_code("dc5ff04c38456b41")
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 torch
import numpy as np

# Analyzing the function:
# - It takes previous_turns which is Union[List[dict], None]
# - It iterates through turns if not None
# - Each turn is a dict with keys that may include task_info keys and other keys
# - For non-task_info keys, it zips their values together
# - Each value in the zip becomes an 'action' dict
# - It accesses turn[key] for keys and action['action_reprs']

# Construct a valid previous_turns argument
previous_turns = [
    {
        'confirmed_task': 'task1',
        'website': 'example.com',
        'domain': 'example',
        'subdomain': 'www',
        'annotation_id': 'annot1',
        'action_reprs': ['click', 'type'],
        'coordinates': [(10, 20), (30, 40)],
        'elements': ['button', 'input']
    },
    {
        'confirmed_task': 'task2',
        'website': 'test.org',
        'domain': 'test',
        'subdomain': 'api',
        'annotation_id': 'annot2',
        'action_reprs': ['scroll', 'hover'],
        'coordinates': [(50, 60), (70, 80)],
        'elements': ['div', 'span']
    }
]

ARGS = {
    'previous_turns': previous_turns
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 2 later) and 5 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
get_descendants 3 other papers
1 dated earlier, 1 the same month, 1 later
2401.01614 (2024-01)
GPT-4V(ision) is a Generalist Web Agent, if Grounded
2402.05930 (2024-02)
WebLINX: Real-World Website Navigation with Multi-Turn Dialogue
2409.11295 (2024-09)
EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage
clean_text 2 other papers
1 dated earlier, 1 later
2401.01614 (2024-01)
GPT-4V(ision) is a Generalist Web Agent, if Grounded
2409.11295 (2024-09)
EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage

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

Web agents powered by Large Language Models (LLMs) have demonstrated remarkable abilities in planning and executing multi-step interactions within complex web-based environments, fulfilling a wide range of web navigation tasks. Despite these advancements, the potential for LLM-powered agents to effectively engage with sequential user instructions in real-world scenarios has not been fully explored. In this work, we introduce a new task of Conversational Web Navigation, which necessitates sophisticated interactions that span multiple turns with both the users and the environment, supported by a specially developed dataset named Multi-Turn Mind2Web (MT-Mind2Web). To tackle the limited context length of LLMs and the context-dependency issue of the conversational tasks, we further propose a novel framework, named selfreflective memory-augmented planning (Self-MAP), which employs memory utilization and self-reflection techniques. Extensive experiments are conducted to benchmark the MT-Mind2Web dataset, and validate the effectiveness of the proposed method. 1

For agents

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

get_harvested_code_for_paper("2402.15057")
get_code_for_paper("2402.15057")
have("2402.15057")

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

curl https://syntology.ai/api/ran/2402.15057.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.15057.svg)](https://syntology.ai/paper/2402.15057)

Connect an agent — have() is free.