SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2108.00648 · 2021

From LSAT: The Progress and Challenges of Complex Reasoning

arXiv · PDF · Open in the Atlas

Code that ran

7 of 11 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 7 have no contract verdict; 0 failures of our instrument; 4 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/2108.00648.json · badge: /api/ran/2108.00648.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
zhongwanjun/AR-LSAT canonical (classifier label)
classifier basis: census
7 ran · 4 no recorded run
FunctionStatusWhere it lives
clean_doc Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/ARM/data_analysis/extract_cp_ner_dp_results.py
code served (permissive licence) · get_code("2771f46777734384")
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

torch.manual_seed(0)

def clean_doc(doc):
    doc = doc.replace('—', ' — ')
    return doc

ARGS = {
    'doc': 'This is a test document with an em dash— in the middle.'
}
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
clean_string Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/LSTM/utils_multiple_choice.py
code served (permissive licence) · get_code("2ee6204e8f1548f9")
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_string(string):
    return re.sub("[^a-zA-Z0-9,.\\'!?]+", ' ', string)

# Construct valid arguments for clean_string
# The function takes a single parameter: string (a string to be cleaned)

ARGS = {
    "string": "Hello, World! This is a test@#$% string with special 123 characters???"
}
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_examples_to_features Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/LSTM/utils_multiple_choice.py
code served (permissive licence) · get_code("92ca0309c5197304")
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
from typing import List

torch.manual_seed(0)

class InputExample(object):
    """A single training/test example for multiple choice"""

    def __init__(self, example_id, contexts, label=None):
        self.example_id = example_id
        self.contexts = contexts
        self.label = label

class InputFeatures(object):

    def __init__(self, example_id, choices_features, label):
        self.example_id = example_id
        self.choices_features = choices_features
        self.label = label

class MockTextField(object):
    """Stand-in for text_field that implements pad and numericalize"""
    
    def __init__(self):
        self.vocab_size = 100
    
    def pad(self, contexts):
        """
        Args:
            contexts: list of str
        Returns:
            tuple of (padded_contexts, lengths)
        """
        # Return mock padded contexts (list of strings) and lengths
        padded_contexts = contexts  # Keep as-is for simplicity
        lengths = torch.tensor([len(c.split()) for c in contexts], dtype=torch.long)
        return padded_contexts, lengths
    
    def numericalize(self, tmp):
        """
        Args:
            tmp: tuple of (padded_contexts, lengths)
        Returns:
            tuple of (input_ids, lengths)
        """
        padded_contexts, lengths = tmp
        num_choices = len(padded_contexts)
        max_len = 8
        # Create mock input_ids tensor
        input_ids = torch.randint(0, self.vocab_size, (num_choices, max_len), dtype=torch.long)
        return input_ids, lengths

# Create mock examples
examples = [
    InputExample(
        example_id='ex_0',
        contexts=['context option 1', 'context option 2', 'context option 3', 'context option 4'],
        label='A'
    ),
    InputExample(
        example_id='ex_1',
        contexts=['another context 1', 'another context 2', 'another context 3', 'another context 4'],
        label='B'
    ),
]

# Create label list
label_list = ['A', 'B', 'C', 'D']

# Create text field
text_field = MockTextField()

# Define ARGS
ARGS = {
    'examples': examples,
    'label_list': label_list,
    'max_length': 128,
    'text_field': text_field
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — 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
extract_program Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/ARM/pipeline/extract_program_argument.py
code served (permissive licence) · get_code("3d29d2111932953c")
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 numpy as np

torch.manual_seed(0)

# The function extract_program(info) takes a single parameter 'info'
# and returns an empty string. Since the function body is minimal,
# 'info' can be any object. We'll use a simple dict as a stand-in
# for a typical info/config object.

info = {
    'name': 'test_program',
    'version': 1,
    'data': [1, 2, 3, 4],
    'model': 'test_model'
}

ARGS = {
    'info': info
}
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 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
find_nearest_ent Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/ARM/pipeline/answer_question_by_tree.py
code served (permissive licence) · get_code("9d60ecb44ac0bbd6")
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

torch.manual_seed(0)

ARGS = {
    "kws": ["what", "is"],
    "question": "what is the capital of france",
    "rows": ["capital", "france", "country"],
    "columns": ["paris", "london", "berlin"]
}
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
find_wspan Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/ARM/pipeline/extract_program_argument.py
code served (permissive licence) · get_code("0316409042f52cb7")
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 int; 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

torch.manual_seed(0)

# Analyzing the function find_wspan:
# - text: a string that is searched in tokens
# - tokens: a list of strings that are checked against text
# - cspan: a tuple/list with two elements (start, end) representing character span positions

# Create simple test data
text = "hello world"
tokens = ["hello", "world", "foo"]
cspan = (0, 11)  # character span covering "hello world"

ARGS = {
    "text": text,
    "tokens": tokens,
    "cspan": cspan
}
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 tuple, ndim 1, dtype family int — 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
rewrite Ran
ran on a synthesized input
zhongwanjun/AR-LSAT/ARM/pipeline/modify_option.py
code served (permissive licence) · get_code("e6cacdf17fcf2a13")
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

torch.manual_seed(0)

# parts: a list of strings that may contain 'and'
parts = ['red and shiny', 'blue', 'green and sparkly']

# rewrite_list: a list of strings with the same length as parts
rewrite_list = ['color', 'shade', 'tone']

ARGS = {
    'parts': parts,
    'rewrite_list': rewrite_list
}
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, 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
choose_question_type Not yet run
unverified (harvested, no recorded run)
zhongwanjun/AR-LSAT/ARM/pipeline/answer_question_by_tree.py
code served (permissive licence) · get_code("305cce488c60e19d")
How it ran

no invocation recorded.

same Not yet run
unverified (harvested, no recorded run)
zhongwanjun/AR-LSAT/ARM/pipeline/extract_program_argument.py
code served (permissive licence) · get_code("5532594d3fb0fa71")
How it ran

no invocation recorded.

select_field Not yet run
unverified (harvested, no recorded run)
zhongwanjun/AR-LSAT/LSTM/main_large.py
pointer only (licence: MIT); the stored copy get_code serves, from yuweihao/reclor, is pointer only (licence: NONE) · get_code("0a546b305d274996")
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 yuweihao/reclor, 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. No invocation recorded.

simple_accuracy Not yet run
unverified (harvested, no recorded run)
zhongwanjun/AR-LSAT/LSTM/main_large.py
code served (permissive licence) · get_code("3c241ecfe3749a6d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 later) and 9 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
simple_accuracy 59 other papers
34 dated earlier, 25 later
1506.05865 (2015-06)
LCSTS: A Large Scale Chinese Short Text Summarization Dataset
1509.01626 (2015-09)
Character-level Convolutional Networks for Text Classification *
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
and 56 more not named here
select_field 9 other papers
5 dated earlier, 4 later
1908.05739 (2019-08)
Abductive Commonsense Reasoning
2002.04326 (2020-02)
ReClor: A Reading Comprehension Dataset Requiring Logical Reasoning
2004.04494 (2020-04)
MuTual: A Dataset for Multi-Turn Dialogue Reasoning
and 6 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

Complex reasoning aims to draw a correct inference based on complex rules. As a hallmark of human intelligence, it involves a degree of explicit reading comprehension, interpretation of logical knowledge and complex rule application. In this paper, we take a step forward in complex reasoning by systematically studying the three challenging and domain-general tasks of the Law School Admission Test (LSAT), including analytical reasoning, logical reasoning and reading comprehension. We propose a hybrid reasoning system to integrate these three tasks and achieve impressive overall performance on the LSAT tests. The experimental results demonstrate that our system endows itself a certain complex reasoning ability, especially the fundamental reading comprehension and challenging logical reasoning capacities. Further analysis also shows the effectiveness of combining the pre-trained models with the task-specific reasoning module, and integrating symbolic knowledge into discrete interpretable reasoning steps in complex reasoning. We further shed a light on the potential future directions, like unsupervised symbolic knowledge extraction, model interpretability, few-shot learning and comprehensive benchmark for complex reasoning.

For agents

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

get_harvested_code_for_paper("2108.00648")
get_code_for_paper("2108.00648")
have("2108.00648")

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

curl https://syntology.ai/api/ran/2108.00648.json

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

[![Syntology run record](https://syntology.ai/api/ran/2108.00648.svg)](https://syntology.ai/paper/2108.00648)

Connect an agent — have() is free.