SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.08043 · 2024

Authorship Style Transfer with Policy Optimization

arXiv · PDF · Open in the Atlas

Code that ran

12 of 13 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; 12 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/2403.08043.json · badge: /api/ran/2403.08043.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
isi-nlp/astrapop canonical (classifier label)
classifier basis: census
12 ran · 1 no recorded run
FunctionStatusWhere it lives
encode_example Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/train_classifier.py
code served (permissive licence) · get_code("07603ec33561b4eb")
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

torch.manual_seed(0)

# Stand-in tokenizer that mimics HuggingFace tokenizer behavior
class MockTokenizer:
    def __call__(self, text, max_length=512, truncation=True):
        # Return mock token IDs and attention mask
        tokens = [101, 2054, 2003, 102]  # Mock token sequence
        return {
            'input_ids': tokens[:max_length],
            'attention_mask': [1] * len(tokens[:max_length])
        }

# Create example dict with required fields
example = {
    'txt': 'This is a sample text for language identification.',
    'native_lang': 'ENG'  # Will need to adjust since ENG is not in LANG2IDX
}

# Actually, let's use a valid language from LANG2IDX
example = {
    'txt': 'This is a sample text for language identification.',
    'native_lang': 'FRA'
}

tokenizer = MockTokenizer()

ARGS = {
    'example': example,
    'tokenizer': tokenizer
}
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
encode_example_paraphrase Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/reddit/train_sft.py
code served (permissive licence) · get_code("8176a95ae3a201b8")
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 numpy as np

class MockTokenizer:
    def __init__(self):
        self.bos_token = '[BOS]'
        self.eos_token = '[EOS]'
        self.vocab_size = 1000
    
    def __call__(self, text, add_special_tokens=False, max_length=128, truncation=False):
        # Simple mock: split text and create token ids
        words = text.split()
        token_ids = [hash(word) % self.vocab_size for word in words]
        if len(token_ids) > max_length:
            token_ids = token_ids[:max_length]
        return {
            'input_ids': token_ids,
            'attention_mask': [1] * len(token_ids)
        }

example = {
    'src': 'The quick brown fox',
    'tgt': 'A fast brown animal'
}

tokenizer = MockTokenizer()

ARGS = {
    'example': example,
    'tokenizer': tokenizer,
    'max_len': 128
}
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 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
encode_paraphrase_example Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/train_sft.py
code served (permissive licence) · get_code("8910f4e5b5d4bbee")
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
from transformers import AutoTokenizer

# Create a minimal tokenizer stand-in that has the required attributes and methods
class MinimalTokenizer:
    def __init__(self):
        self.bos_token = "[BOS]"
        self.eos_token = "[EOS]"
        self.vocab_size = 1000
    
    def __call__(self, text, add_special_tokens=False, max_length=128, truncation=True):
        # Simple mock tokenization: split by spaces and convert to token ids
        tokens = text.split()
        token_ids = [hash(token) % self.vocab_size for token in tokens]
        
        # Apply truncation if needed
        if truncation and len(token_ids) > max_length:
            token_ids = token_ids[:max_length]
        
        return {
            'input_ids': token_ids,
            'attention_mask': [1] * len(token_ids)
        }

# Create example data
example = {
    'src': 'The quick brown fox',
    'tgt': 'A fast brown fox'
}

# Create tokenizer
tokenizer = MinimalTokenizer()

# Define all arguments
ARGS = {
    'example': example,
    'tokenizer': tokenizer,
    'max_len': 128,
    'direction': 'forward'
}
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 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
encode_transfer_example Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/train_sft.py
code served (permissive licence) · get_code("f463ab3c42a28cb4")
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 numpy as np
from transformers import PreTrainedTokenizer

torch.manual_seed(0)

class MockTokenizer:
    def __init__(self):
        self.bos_token = "[BOS]"
        self.eos_token = "[EOS]"
        self.vocab = {
            "[BOS]": 0,
            "[EOS]": 1,
            "[SRC]": 2,
            "[/SRC]": 3,
            "hello": 4,
            "world": 5,
            "test": 6,
        }
    
    def __call__(self, text, add_special_tokens=False, max_length=1000, truncation=True):
        tokens = text.split()
        token_ids = [self.vocab.get(token, 5) for token in tokens]
        
        if truncation and len(token_ids) > max_length:
            token_ids = token_ids[:max_length]
        
        return {
            'input_ids': token_ids,
            'attention_mask': [1] * len(token_ids)
        }

example = {
    'src': 'hello world',
    'tgt': 'test output'
}

tokenizer = MockTokenizer()

ARGS = {
    'example': example,
    'tokenizer': tokenizer,
    'max_len': 1000
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, transformers — 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
generate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isi-nlp/astrapop/src/ets/generate.py
code served (permissive licence) · get_code("6c1788682bf64080")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.nn as nn

torch.manual_seed(0)

# Stand-in tokenizer that mimics HuggingFace tokenizer behavior
class MockTokenizer:
    def __init__(self):
        self.pad_token_id = 0
    
    def __call__(self, prompts, add_special_tokens=False, max_length=384, 
                 padding=True, truncation=True, return_tensors='pt'):
        # Return mock encoded prompts
        batch_size = len(prompts) if isinstance(prompts, list) else 1
        seq_len = 5
        return {
            'input_ids': torch.randint(1, 100, (batch_size, seq_len)),
            'attention_mask': torch.ones(batch_size, seq_len)
        }
    
    def batch_decode(self, token_ids, skip_special_tokens=True):
        # Return mock decoded strings
        batch_size = token_ids.size(0)
        return [f"generated text {i}" for i in range(batch_size)]

# Stand-in generation config
class MockGenerationConfig:
    def __init__(self):
        self.max_length = 128
        self.num_beams = 1

# Stand-in model with generate method
class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
        self._adapter_name = None
    
    def set_adapter(self, adapter_name):
        self._adapter_name = adapter_name
    
    def generate(self, input_ids, attention_mask, generation_config, max_new_tokens):
        # Return mock generated token ids
        batch_size = input_ids.size(0)
        input_len = input_ids.size(1)
        new_tokens = 3
        # Concatenate input_ids with new generated tokens
        generated = torch.cat([
            input_ids,
            torch.randint(1, 100, (batch_size, new_tokens))
        ], dim=1)
        return generated

# Construct arguments
prompts = ["Hello world", "How are you"]
model = MockModel()
tokenizer = MockTokenizer()
generation_config = MockGenerationConfig()
adapter_name = "test_adapter"
max_new_tokens = 128

ARGS = {
    "prompts": prompts,
    "model": model,
    "tokenizer": tokenizer,
    "generation_config": generation_config,
    "adapter_name": adapter_name,
    "max_new_tokens": max_new_tokens
}
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
generate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
isi-nlp/astrapop/src/ets/generate_dpo_cpo_data.py
code served (permissive licence) · get_code("4fef07692ec2e44f")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.nn as nn
from unittest.mock import Mock

torch.manual_seed(0)

# Minimal tokenizer stand-in
class MinimalTokenizer:
    def __call__(self, prompts, add_special_tokens=False, max_length=1000, 
                 padding=True, truncation=True, return_tensors='pt'):
        batch_size = len(prompts) if isinstance(prompts, list) else 1
        return {
            'input_ids': torch.randint(0, 100, (batch_size, 5)),
            'attention_mask': torch.ones(batch_size, 5, dtype=torch.long)
        }
    
    def batch_decode(self, token_ids, skip_special_tokens=True):
        return ["generated text " + str(i) for i in range(token_ids.size(0))]

# Minimal generation config
class MinimalGenerationConfig:
    pass

# Minimal model with generate method
class MinimalModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
    
    def generate(self, input_ids, attention_mask, generation_config, 
                 max_new_tokens, refs_input_ids=None, refs_attn_mask=None):
        batch_size = input_ids.size(0)
        seq_len = input_ids.size(1)
        new_tokens = torch.randint(0, 100, (batch_size, max_new_tokens))
        return torch.cat([input_ids, new_tokens], dim=1)
    
    def set_adapter(self, adapter_name):
        pass

# Create instances
tokenizer = MinimalTokenizer()
model = MinimalModel()
generation_config = MinimalGenerationConfig()
prompts = ["Hello world", "Test prompt"]

# Optional encoded_refs for testing the refs path
encoded_refs = {
    'input_ids': torch.randint(0, 100, (2, 6)),
    'attention_mask': torch.ones(2, 6, dtype=torch.long)
}

ARGS = {
    'prompts': prompts,
    'model': model,
    'tokenizer': tokenizer,
    'generation_config': generation_config,
    'adapter_name': None,
    'max_new_tokens': 128,
    'encoded_refs': None
}
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
get_concat_transfer_prompt Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/reddit/train_ppo.py
code served (permissive licence) · get_code("44f9b5a95924386d")
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)

# Analyze the function:
# - get_concat_transfer_prompt takes two parameters: src and refs
# - src is used in an f-string: f'[SRC]{src}[/SRC]'
# - refs is iterated over in a list comprehension: [f'[REF]{ref}[/REF]' for ref in refs]
# - Both src and refs are converted to strings in f-strings
# - So they can be strings or any objects with __str__ method

src = "This is the source text"
refs = ["Reference one", "Reference two", "Reference three"]

ARGS = {
    "src": src,
    "refs": refs,
}
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_paraphrase_prompt Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/generate.py
code served (permissive licence) · get_code("9cc59215e82b53ec")
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)

src = "The quick brown fox jumps over the lazy dog"

ARGS = {
    "src": src
}
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_random_tgt_lang Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/train_ppo.py
code served (permissive licence) · get_code("9d3e98497d85a686")
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 random

LANG2IDX = {'ARA': 0, 'DEU': 1, 'FRA': 2, 'HIN': 3, 'ITA': 4, 'JPN': 5, 'KOR': 6, 'SPA': 7, 'TEL': 8, 'TUR': 9, 'ZHO': 10}

def get_random_tgt_lang(src_lang):
    langs = sorted(LANG2IDX.keys())
    langs.remove(src_lang)
    random_tgt_lang = random.choice(langs)
    return random_tgt_lang

random.seed(0)

ARGS = {
    'src_lang': 'ARA'
}
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
get_styll_prompt Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/reddit/train_sft.py
code served (permissive licence) · get_code("724242a144261aac")
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

# Define ARGS for get_styll_prompt
# The function takes:
# - src: a string (the source text to be rewritten)
# - refs: a list of tuples, where each tuple contains (paraphrased_tgt_text, tgt_text)

src = "The quick brown fox jumps over the lazy dog"

refs = [
    ("A fast brown fox leaps over a sleepy dog", "The speedy fox jumped over the sleeping dog"),
    ("The rapid fox hops across the inactive dog", "A quick fox vaulted over the lethargic dog"),
]

ARGS = {
    "src": src,
    "refs": refs,
}
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_transfer_prompt Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/generate.py
code served (permissive licence) · get_code("af515ba22c9dfe28")
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)

ARGS = {
    "src": "Hello, how are you?",
    "native_lang": "en"
}
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_transfer_prompt Ran
ran; our contract draft was wrong, not the code
isi-nlp/astrapop/src/ets/generate_dpo_cpo_data.py
code served (permissive licence) · get_code("ec477fe8b167e730")
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 get_transfer_prompt(src):
    prompt = f'[SRC]{src}[/SRC]'
    return prompt

ARGS = {
    'src': 'This is a source text for transfer learning.'
}
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
encode_example_transfer Not yet run
unverified (harvested, no recorded run)
isi-nlp/astrapop/src/reddit/train_sft.py
code served (permissive licence) · get_code("c42a75435bdfbae3")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 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 13 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

Authorship style transfer aims to rewrite a given text into a specified target while preserving the original meaning in the source. Existing approaches rely on the availability of a large number of target style exemplars for model training. However, these overlook cases where a limited number of target style examples are available. The development of parameter-efficient transfer learning techniques and policy optimization (PO) approaches suggest lightweight PO is a feasible approach to low-resource style transfer. In this work, we propose a simple two-stage tune-and-optimize technique for low-resource textual style transfer. We apply our technique to authorship transfer as well as a larger-data native language style task and in both cases find it outperforms state-of-the-art baseline models.

For agents

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

get_harvested_code_for_paper("2403.08043")
get_code_for_paper("2403.08043")
have("2403.08043")

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

curl https://syntology.ai/api/ran/2403.08043.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.08043.svg)](https://syntology.ai/paper/2403.08043)

Connect an agent — have() is free.