SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.15898 · 2026

MultiCube-RAG for Multi-hop Question Answering

Wei Hu, Jiawei Han, Jin Bowen, Jimeng Shi, Sizhe Zhou, Runchu Tian, Seongku Kang, Wonbin Kweon, Shaowen Wang, Dingqi Ye, Yunfan Kang, Multicube

arXiv · PDF · Open in the Atlas

Code that ran

5 of 10 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, 2 violated it and 0 have no contract verdict; 3 failures of our instrument; 5 no recorded run), measured 2026-11-22; 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/2602.15898.json · badge: /api/ran/2602.15898.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
OpenBMB/UltraRAG canonical (classifier label)
classifier basis: census
5 ran · 5 no recorded run
FunctionStatusWhere it lives
clean_text Ran
ran; our contract draft was wrong, not the code
OpenBMB/UltraRAG/servers/corpus/src/corpus.py
code served (permissive licence) · get_code("4de9e8ee4cbe4329")
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: str) -> str:
    """Clean text by normalizing whitespace and line breaks.

    Args:
        text: Input text to clean

    Returns:
        Cleaned text string
    """
    if not text:
        return ''
    text = text.replace('\u3000', ' ')
    text = text.replace('\r\n', '\n').replace('\r', '\n')
    text = re.sub('\\n{3,}', '\n\n', text)
    return text.strip()


ARGS = {
    'text': 'Hello\u3000world\r\nThis is a test\n\n\n\nMultiple lines'
}
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
elem_match Ran
ran, violated the contract we drafted
OpenBMB/UltraRAG/src/ultrarag/client.py
code served (permissive licence) · get_code("54ae30263c24b31b")
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
import numpy as np
from typing import Dict, List, Tuple

torch.manual_seed(0)

# Construct elem dictionary with branch state keys
elem = {
    'branch0_state': 'active',
    'branch1_state': 'inactive',
    'branch2_state': 'pending',
    'other_key': 'some_value'
}

# Construct pairs list with (depth, state) tuples
pairs = [
    (0, 'active'),
    (1, 'inactive'),
    (2, 'pending')
]

ARGS = {
    'elem': elem,
    'pairs': pairs
}
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 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
escape_html Ran
ran; our contract draft was wrong, not the code
OpenBMB/UltraRAG/script/case_study.py
code served (permissive licence) · get_code("10e5a6bb74a0eb89")
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

ARGS = {
    's': '<div>Hello & goodbye</div>'
}
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
is_wrapped_list Ran
ran, violated the contract we drafted
OpenBMB/UltraRAG/src/ultrarag/client.py
code served (permissive licence) · get_code("2fb48b1aafbd7600")
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
import torch.nn as nn

torch.manual_seed(0)

# Test case 1: Valid wrapped list format with branch_state
wrapped_list_valid = [
    {
        "data": torch.randn(2, 4),
        "branch0_state": {"param": 1},
        "branch1_state": {"param": 2}
    }
]

# Test case 2: Empty list (should return False)
empty_list = []

# Test case 3: List with dict but no branch*_state keys (should return False)
non_wrapped_list = [
    {
        "data": torch.randn(2, 4),
        "other_key": "value"
    }
]

# Test case 4: List with non-dict first element (should return False)
list_with_non_dict = [1, 2, 3]

# Test case 5: Not a list (should return False)
not_a_list = "not a list"

# Use the wrapped_list_valid as the primary test case
ARGS = {
    "lst": wrapped_list_valid
}
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
parse_path Ran
ran; our contract draft was wrong, not the code
OpenBMB/UltraRAG/src/ultrarag/client.py
code served (permissive licence) · get_code("5645318660d069dd")
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 2, 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

# Analyze the function:
# - parse_path takes a single parameter: path (str)
# - It parses a branch path string into list of (depth, state) tuples
# - The function handles ROOT constant and SEP separator
# - It splits the path and extracts depth and state from segments like "branch1_finished"

ARGS = {
    'path': 'BASE/branch1_finished/branch2_retry/branch3_error'
}
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 2, 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
PipelineCall Not yet run
unverified (harvested, no recorded run)
OpenBMB/UltraRAG/src/ultrarag/api.py
code served (permissive licence) · get_code("82b55ef05c61a161")
How it ran

no invocation recorded.

get_logger Not yet run
unverified (harvested, no recorded run)
OpenBMB/UltraRAG/src/ultrarag/mcp_logging.py
code served (permissive licence) · get_code("9e03092ffc8351c2")
How it ran

no invocation recorded.

load_cases Not yet run
unverified (harvested, no recorded run)
OpenBMB/UltraRAG/script/case_study.py
code served (permissive licence) · get_code("8ef4d0e3d5ff2a82")
How it ran

no invocation recorded.

popen_follow_parent Not yet run
unverified (harvested, no recorded run)
OpenBMB/UltraRAG/src/ultrarag/utils.py
code served (permissive licence) · get_code("ab7464e11cc77525")
How it ran

no invocation recorded.

reflow_paragraphs Not yet run
unverified (harvested, no recorded run)
OpenBMB/UltraRAG/servers/corpus/src/corpus.py
code served (permissive licence) · get_code("67e431f4417e4cb8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 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 10 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 2026-11-22.

Repositories linked to this paper

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

Abstract

Multi-hop question answering (QA) necessitates multi-step reasoning and retrieval across interconnected subjects, attributes, and relations. Existing retrieval-augmented generation (RAG) methods struggle to capture these structural semantics accurately, resulting in suboptimal performance. Graph-based RAGs structure such information in graphs, but the resulting graphs are often noisy and computationally expensive. Moreover, most methods rely on singlestep retrieval, neglecting the need for multi-hop reasoning processes. Recent training-based approaches attempt to incentivize the large language models (LLMs) for iterative reasoning and retrieval, but their training processes are prone to unstable convergence and high computational overhead. To address these limitations, we devise an ontology-based cube structure with multiple and orthogonal dimensions to model structural subjects, attributes, and relations. Built on the cube structure, we propose MultiCube-RAG, a training-free method consisting of multiple cubes for multi-step reasoning and retrieval. Each cube specializes in modeling a class of subjects, so that MultiCube-RAG flexibly selects the most suitable cubes to acquire the relevant knowledge precisely. To enhance the query-based reasoning and retrieval, our method decomposes a complex multi-hop query into a set of simple subqueries along cube dimensions and conquers each of them sequentially. Experiments on four multi-hop QA datasets show that MultiCube-RAG improves response accuracy by 8.9% over the average performance of various baselines. Notably, we also demonstrate that our method performs with greater efficiency and inherent explainability. The code is available at https:// anonymous.4open.science/ r/ CubeRAG/ .

For agents

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

get_harvested_code_for_paper("2602.15898")
get_code_for_paper("2602.15898")
have("2602.15898")

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

curl https://syntology.ai/api/ran/2602.15898.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.15898.svg)](https://syntology.ai/paper/2602.15898)

Connect an agent — have() is free.