SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2504.01698 · 2025

Do Theory of Mind Benchmarks Need Explicit Human-like Reasoning in Language Models?

arXiv · PDF · Open in the Atlas

Code that ran

11 of 15 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; 1 honoured the contract we drafted, 1 violated it and 6 have no contract verdict; 3 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/2504.01698.json · badge: /api/ran/2504.01698.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
bigai-ai/ToM-RL canonical (classifier label)
classifier basis not recorded on the link
11 ran · 4 no recorded run
FunctionStatusWhere it lives
create_huggingface_actor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
bigai-ai/ToM-RL/verl/utils/model.py
code served (permissive licence) · get_code("360c671d6041ba3b")
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 MockCausalLMModel. 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
from unittest.mock import Mock, patch
from transformers import AutoConfig, PretrainedConfig

torch.manual_seed(0)

# Create a minimal mock config class
class MockPretrainedConfig(PretrainedConfig):
    model_type = "mock_model"
    
    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self.hidden_size = 64
        self.num_hidden_layers = 2
        self.num_attention_heads = 2
        self.intermediate_size = 128
        self.vocab_size = 1000
        self.max_position_embeddings = 512

# Create a minimal mock model
class MockCausalLMModel(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.config = config
        self.embeddings = nn.Embedding(config.vocab_size, config.hidden_size)
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size)
    
    def forward(self, input_ids):
        x = self.embeddings(input_ids)
        return self.lm_head(x)

# Mock AutoConfig.from_pretrained to return our mock config
def mock_from_pretrained(model_name, trust_remote_code=False):
    return MockPretrainedConfig()

# Mock AutoModelForCausalLM.from_config to return our mock model
def mock_from_config(config, **kwargs):
    return MockCausalLMModel(config)

# Set up patches
import transformers
original_from_pretrained = transformers.AutoConfig.from_pretrained
original_from_config = transformers.AutoModelForCausalLM.from_config

transformers.AutoConfig.from_pretrained = staticmethod(mock_from_pretrained)
transformers.AutoModelForCausalLM.from_config = staticmethod(mock_from_config)

ARGS = {
    "model_name": "gpt2",
    "override_config_kwargs": {"hidden_size": 128, "num_hidden_layers": 4},
    "automodel_kwargs": {"trust_remote_code": False}
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MockCausalLMModel — a MockCausalLMModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockCausalLMModel 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
extract_solution Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom.py
code served (permissive licence) · get_code("1f0645c6a0d2044e")
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
from typing import Dict, Tuple, Optional, Union, Any

torch_manual_seed_set = False

# Test case 1: Response with "Assistant:" header and answer tags
solution_str_1 = """Some preamble here
Assistant: Let me solve this step by step.
First, I calculate something.
Then I get the result.
<answer>42</answer>
Some trailing text after answer."""

# Test case 2: Response with "<|im_start|>assistant" header and answer tags
solution_str_2 = """<|im_start|>system
You are a helpful assistant.
<|im_start|>assistant
Working through the problem...
The calculation shows:
<answer>The final solution is 3.14159</answer>
More text here."""

# Test case 3: Response with multiple answer tags (should return the last one)
solution_str_3 = """Assistant: First attempt: <answer>wrong</answer>
Let me recalculate.
<answer>correct answer</answer>"""

# Test case 4: Response without assistant header (should return None for answer)
solution_str_4 = """Just some random text
<answer>this won't be extracted</answer>"""

# Test case 5: Response with Assistant header but no answer tags
solution_str_5 = """Assistant: I'm solving this problem but forgot to include answer tags.
The answer is somewhere in here but not properly tagged."""

# Select one for the main test - using test case 1
ARGS = {
    'solution_str': solution_str_1
}
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
extract_step Ran
ran, honoured the contract we drafted
bigai-ai/ToM-RL/verl/trainer/fsdp_sft_trainer.py
code served (permissive licence) · get_code("216e28c040173a61")
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) and returned int, ndim 0, dtype family int, 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 re

def extract_step(path):
    match = re.search('global_step_(\\d+)', path)
    if match:
        return int(match.group(1))
    return None

# Test case 1: path with global_step pattern
path_with_step = "/models/checkpoint_global_step_42/model.pt"

# Test case 2: path without global_step pattern
path_without_step = "/models/checkpoint_final/model.pt"

ARGS = {
    "path": path_with_step
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 int, ndim 0, dtype family int, 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
extract_xml_answer Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom_2.py
code served (permissive licence) · get_code("55d752fd12d893c2")
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 extract_xml_answer(text: str) -> str:
    answer_pattern = '<answer>(.*?)</answer>'
    matches = list(re.finditer(answer_pattern, text, re.DOTALL))
    if not matches:
        return None
    final_answer = matches[-1].group(1).strip()
    return final_answer

# Setup code
text_with_single_answer = "Some text <answer>First Answer</answer> more text"
text_with_multiple_answers = "Text <answer>Answer 1</answer> middle <answer>Final Answer</answer> end"
text_without_answer = "Just some plain text with no XML tags"

ARGS = {
    "text": text_with_multiple_answers
}
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
is_transformers_version_in_range Ran
ran, violated the contract we drafted
bigai-ai/ToM-RL/verl/models/transformers/monkey_patch.py
code served (permissive licence) · get_code("42a7954a8cfa89e0")
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 packaging import version

# The function checks if the transformers package version falls within a given range
# It uses version.parse() to compare version strings
# We need to provide valid version strings as arguments

ARGS = {
    'min_version': '4.0.0',
    'max_version': '5.0.0',
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, packaging, 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
normalize_answer Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom.py
code served (permissive licence) · get_code("1436322b957b9b34")
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 normalize_answer(answer: str) -> str:
    """Normalizes the answer text for better comparison.
    
    Args:
        answer: Raw answer text
        
    Returns:
        Normalized answer text
    """
    normalized = answer.lower()
    normalized = re.sub('\\s+', ' ', normalized).strip()
    normalized = re.sub('[.,;:!?]', '', normalized)
    normalized = re.sub('_', ' ', normalized)
    return normalized


ARGS = {
    "answer": "Hello, World! This is a test_answer with  multiple   spaces."
}
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
normalize_answer Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom_2.py
code served (permissive licence) · get_code("c5322e5a65a9bea1")
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 normalize_answer(answer: str) -> str:
    """Normalizes the answer text for better comparison.
    Args:
        answer: Raw answer text
    Returns:
        Normalized answer text
    """
    normalized = answer.lower()
    normalized = re.sub('\\s+', ' ', normalized).strip()
    normalized = re.sub('[.,;:!?]', '', normalized)
    return normalized


ARGS = {
    "answer": "Hello, World!  This is a TEST sentence."
}
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
reward_func Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom_2.py
code served (permissive licence) · get_code("a5c20c9ff16e297d")
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 int, ndim 0, dtype family int, 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 re

def extract_xml_answer(text: str) -> str:
    answer_pattern = '<answer>(.*?)</answer>'
    matches = list(re.finditer(answer_pattern, text, re.DOTALL))
    if not matches:
        return None
    final_answer = matches[-1].group(1).strip()
    return final_answer

def normalize_answer(answer: str) -> str:
    """Normalizes the answer text for better comparison.
    Args:
        answer: Raw answer text
    Returns:
        Normalized answer text
    """
    normalized = answer.lower()
    normalized = re.sub('\\s+', ' ', normalized).strip()
    normalized = re.sub('[.,;:!?]', '', normalized)
    return normalized

def reward_func(response, answer):
    pattern = '.*?<answer>.*?</answer>\\s*\\b<|im_end|>$'
    tags = {'ans_start': ('<answer>', 1), 'ans_end': ('</answer>', 1)}
    counts = 0
    for tag_name, (tag_str, expected_count) in tags.items():
        count = response.count(tag_str)
        if count == expected_count:
            counts += 1
    if counts == 2:
        match = re.match(pattern, response, re.DOTALL | re.MULTILINE)
        if match:
            response_ = extract_xml_answer(response)
            norm_response = normalize_answer(response_)
            norm_answer = normalize_answer(answer)
            ans_pattern = "\\b(?:in|at|on|inside|)?\\s*(?:the\\s*)?(?:\\w+'s\\s*)?" + re.escape(norm_answer) + '\\s*\\b$'
            match = re.match(ans_pattern, norm_response, re.DOTALL | re.MULTILINE)
            if match:
                print(f'Right format and exactly match, score: 2, response: ({norm_response}), answer: ({norm_answer})')
                return 2
            else:
                print(f'Right format but wrong answer, score: 0, response: ({norm_response}), answer: ({norm_answer})')
                return 0
        else:
            print('Right format (tag counts correct) but no answer found, score: 0')
            return 0
    print(f'Wrong format, score: 0')
    return 0

ARGS = {
    "response": "<answer>the kitchen</answer> <|im_end|>",
    "answer": "kitchen"
}
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 int, ndim 0, dtype family int, 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
union_tensor_dict Ran
ran; our contract draft was wrong, not the code
bigai-ai/ToM-RL/verl/protocol.py
code served (permissive licence) · get_code("21331a58f93375e4")
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 TensorDict. 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 tensordict import TensorDict

torch.manual_seed(0)

# Create two TensorDicts with identical batch size
batch_size = torch.Size([2])

# First tensor dict with some keys
tensor_dict1 = TensorDict(
    {
        "shared_key": torch.randn(2, 4),
        "key1": torch.randn(2, 3),
    },
    batch_size=batch_size,
)

# Second tensor dict with overlapping and new keys
tensor_dict2 = TensorDict(
    {
        "shared_key": tensor_dict1["shared_key"].clone(),  # Same tensor values
        "key2": torch.randn(2, 5),
    },
    batch_size=batch_size,
)

ARGS = {
    "tensor_dict1": tensor_dict1,
    "tensor_dict2": tensor_dict2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensordict, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TensorDict — a TensorDict 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
unpad_dataproto Ran
ran; our contract draft was wrong, not the code
bigai-ai/ToM-RL/verl/protocol.py
code served (permissive licence) · get_code("25f0ea3f460f6ce1")
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) and returned DataProto. 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)

class DataProto:
    """Stand-in class for DataProto that supports slicing."""
    def __init__(self, data):
        self.data = data
    
    def __getitem__(self, key):
        """Support slicing to return a new DataProto with sliced data."""
        sliced_data = self.data[key]
        return DataProto(sliced_data)
    
    def __repr__(self):
        return f"DataProto({self.data})"

# Create a small DataProto instance with some data
data_tensor = torch.randn(2, 4)
data_proto = DataProto(data_tensor)

ARGS = {
    "data": data_proto,
    "pad_size": 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataProto — a DataProto output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataProto 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
validate_response_structure Ran
ran on a synthesized input
bigai-ai/ToM-RL/verl/utils/reward_score/explore_tom.py
code served (permissive licence) · get_code("c53a02ef0eee0abe")
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 response structure with all tags properly ordered
valid_response = "<think>This is my reasoning process</think><answer>This is my final answer</answer>"

# Test case 2: Response with tags but in wrong order (for testing validation failure)
invalid_order_response = "<answer>Wrong order</answer><think>Should be first</think>"

# Test case 3: Response missing a closing tag
missing_tag_response = "<think>No closing think tag<answer>Missing end</answer>"

# Test case 4: Response with duplicate tags
duplicate_tags_response = "<think>First think</think><think>Second think</think><answer>Answer</answer>"

# For the main execution, use a valid response
ARGS = {
    "processed_str": "<think>Let me think about this problem step by step. First I'll analyze the input, then work through the logic.</think><answer>The answer to the problem is that we need to validate the structure carefully.</answer>"
}
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
apply_monkey_patch Not yet run
unverified (harvested, no recorded run)
bigai-ai/ToM-RL/verl/models/transformers/monkey_patch.py
code served (permissive licence) · get_code("1fb4adb621eefddc")
How it ran

no invocation recorded.

get_huggingface_actor_config Not yet run
unverified (harvested, no recorded run)
bigai-ai/ToM-RL/verl/utils/model.py
code served (permissive licence) · get_code("19472d4343d0a082")
How it ran

no invocation recorded.

get_weight_loader Not yet run
unverified (harvested, no recorded run)
bigai-ai/ToM-RL/verl/models/weight_loader_registry.py
code served (permissive licence) · get_code("919cf310ad4c189a")
How it ran

no invocation recorded.

squeeze Not yet run
unverified (harvested, no recorded run)
bigai-ai/ToM-RL/verl/utils/model.py
code served (permissive licence) · get_code("3b15e2ac7497c441")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 9 are also present in other papers whose code we hold (9 with at least one paper dated earlier, 9 the same month, 9 later) and 6 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
unpad_dataproto 89 other papers
7 dated earlier, 4 the same month, 74 later, 1 same year, order not determinable, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 86 more not named here
squeeze 82 other papers
7 dated earlier, 3 the same month, 68 later, 1 same year, order not determinable, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 79 more not named here
union_tensor_dict 64 other papers
7 dated earlier, 4 the same month, 49 later, 1 same year, order not determinable, 3 undated · 4 only as an assembly brick
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 61 more not named here
get_huggingface_actor_config 53 other papers
7 dated earlier, 3 the same month, 40 later, 1 same year, order not determinable, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 50 more not named here
get_weight_loader 20 other papers
3 dated earlier, 2 the same month, 13 later, 1 same year, order not determinable, 1 undated
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
2502.14768 (2025-02)
Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
and 17 more not named here
extract_step 19 other papers
2 dated earlier, 1 the same month, 14 later, 1 same year, order not determinable, 1 undated
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
2502.14768 (2025-02)
Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
and 16 more not named here
create_huggingface_actor 17 other papers
2 dated earlier, 1 the same month, 12 later, 1 same year, order not determinable, 1 undated
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
2502.14768 (2025-02)
Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
2503.10460 (2025-03)
Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond
and 14 more not named here
is_transformers_version_in_range 14 other papers
3 dated earlier, 2 the same month, 7 later, 1 same year, order not determinable, 1 undated
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
2502.14768 (2025-02)
Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
and 11 more not named here
apply_monkey_patch 10 other papers
2 dated earlier, 1 the same month, 5 later, 1 same year, order not determinable, 1 undated
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
2502.14768 (2025-02)
Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
and 7 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

Theory of Mind (ToM), the ability to attribute mental states to others, is fundamental for human social intelligence and a critical capability for advanced Artificial Intelligence. Recent advancements in Large Language Models (LLMs) have shown promising performance on ToM benchmarks, raising the question: Do these benchmarks necessitate explicit human-like reasoning processes, or can models succeed through alternative strategies? We investigate this question empirically by applying Reinforcement Learning (RL) and Supervised Fine-Tuning (SFT) to LLMs of varying scales (0.5B to 7B parameters) and evaluating them across multiple ToM datasets. Our results reveal a scale-dependent impact of RL: while RL significantly improves accuracy and fosters high-quality, interpretable, and transferable belief-tracking reasoning in larger models (7B), it leads to "reasoning collapse" in smaller models ($\leq$3B), where high accuracy and generalization ability are achieved via drastically shortened, less meaningful responses. Surprisingly, further SFT achieves competitive and generalizable performance across these benchmarks, often matching or exceeding RL models in accuracy, despite not being explicitly trained to produce structured reasoning traces. These findings highlight a critical discrepancy between benchmark accuracy and the nature of learned reasoning. Our work suggests that current ToM benchmarks may be solvable without requiring the explicit, human-like simulation of mental states they were designed to probe. LLMs, particularly when scale is limited or training signals focus solely on output correctness, may leverage alternative rules effective for benchmark data structures.

For agents

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

get_harvested_code_for_paper("2504.01698")
get_code_for_paper("2504.01698")
have("2504.01698")

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

curl https://syntology.ai/api/ran/2504.01698.json

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

[![Syntology run record](https://syntology.ai/api/ran/2504.01698.svg)](https://syntology.ai/paper/2504.01698)

Connect an agent — have() is free.