Papers › MGTBench: Benchmarking Machine-Generated Text Detection

MGTBench: Benchmarking Machine-Generated Text Detection

26 Mar 2023arXiv:2303.14822archive 2025-07-28

Xinlei He, Xinyue Shen, Zeyuan Chen, Michael Backes, Yang Zhang

Nowadays, powerful large language models (LLMs) such as ChatGPT have demonstrated revolutionary power in a variety of tasks. Consequently, the detection of machine-generated texts (MGTs) is becoming increasingly crucial as LLMs become more advanced and prevalent. These models have the ability to generate human-like language, making it challenging to discern whether a text is authored by a human or a machine. This raises concerns regarding authenticity, accountability, and potential bias. However, existing methods for detecting MGTs are evaluated using different model architectures, datasets, and experimental settings, resulting in a lack of a comprehensive evaluation framework that encompasses various methodologies. Furthermore, it remains unclear how existing detection methods would perform against powerful LLMs. In this paper, we fill this gap by proposing the first benchmark framework for MGT detection against powerful LLMs, named MGTBench. Extensive evaluations on public datasets with curated texts generated by various powerful LLMs such as ChatGPT-turbo and Claude demonstrate the effectiveness of different detection methods. Our ablation study shows that a larger number of words in general leads to better performance and most detection methods can achieve similar performance with much fewer training samples. Moreover, we delve into a more challenging task: text attribution. Our findings indicate that the model-based detection methods still perform well in the text attribution task. To investigate the robustness of different detection methods, we consider three adversarial attacks, namely paraphrasing, random spacing, and adversarial perturbations. We discover that these attacks can significantly diminish detection effectiveness, underscoring the critical need for the development of more robust detection methods.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

No sample from the archive's official repository (xinleihe/mgtbench) is linked to this paper; the 26 samples below come from 3 community repositories. Syntology harvested code from xinleihe/mgtbench for another paper.

xinleihe/mgtbench: Syntology's graph links these samples to ACL Anthology id 2023.emnlp-main.616 (that paper's own run record: 12 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method.

Syntology 26 code samples harvested from 3 repositories linked to this paper: 19 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 7 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 6 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 26 samples from 3 repositories, 19 ran (12 with no instrument failure; 7 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

xinleihe/mgtbench officialmentioned in papermentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to ACL Anthology id 2023.emnlp-main.616 (that paper's own run record: 12 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report
kinit-sk/imgtb mentioned on GitHubpytorchSyntology: 1 sample, 1 ran (1 instrument-failed) · 0 unverified report
michalspiegel/imgtb mentioned on GitHubpytorchMITSyntology: 16 samples, 11 ran (4 instrument-failed) · 5 unverified report
trustairlab/mgtbench mentioned on GitHubpytorchMITSyntology: 9 samples, 7 ran (2 instrument-failed) · 2 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

26 samples harvested: 19 ran (of which 0 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 7 where Syntology's instrument failed) · 7 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

6ran · our draft was wrong
1ran; the contract check could not run
12ran
7unverified

Licence: 5 of the 26 samples are pointer only, meaning Syntology does not serve that copy's text. For 5 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

kinit-sk/imgtb community (archive-listed) 1 sample, 1 ran (1 instrument-failed) · 0 unverified
run_all_available benchmark.py ran · our draft was wrong MIT (permissive) · c032d17914012abc · report fixture: signature data:SimpleData/0/?, method_config:dict/0/?, available_experiments:list/1/callable · returned dict · level recorded 2026-09-22
michalspiegel/imgtb community (archive-listed) 16 samples, 11 ran (4 instrument-failed) · 5 unverified
filter_test_data methods/utils.py ran MIT (permissive) · 3409c93cd092820c · report fixture: signature data:dict/0/?, max_length:int/0/int · returned dict · level recorded 2026-09-25
get_supervised_model_prediction_multi_classes methods/abstract_methods/supervised_experiment.py ran MIT (permissive) here; the stored copy is pointer only · 68c8b968f55907b7 · report fixture: signature model:callable/0/?, tokenizer:callable/0/?, data:list/1/str, batch_size:int/0/int, DEVICE:str/0/str, pos_bit:int/0/int · returned list · level recorded 2026-09-24
load_TruthfulQA lib/dataset_loader_attribution.py ran MIT (permissive) · 2eeb7e1b194357da · report fixture: signature cache_dir:str/0/str · returned dict · level recorded 2026-09-25
make_method_names_unique results_analysis.py ran MIT (permissive) · 8709358702e01ea5 · report fixture: signature results:dict/0/? · returned dict · level recorded 2026-09-25
process_spaces in 6 papers lib/dataset_loader_attribution.py ran · our draft was wrong fingerprinted MIT (permissive) · c5d3444c3ea71db9 · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-26
process_text_truthfulqa_adv lib/dataset_loader_attribution.py ran · our draft was wrong fingerprinted MIT (permissive) · dd7573db30a9bd4e · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-26
read_multiple_to_pandas lib/dataset_loader.py ran MIT (permissive) · d4bc8c999e20811c · report fixture: signature datasets_params:list/1/dict, is_interactive:bool/0/bool · returned dict · level recorded 2026-09-24
select_train_data methods/utils.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 416e7dabab64e070 · report fixture: signature data:dict/0/?, select_num:int/0/int · returned dict · level recorded 2026-09-24
set_pos_bit methods/abstract_methods/supervised_experiment.py ran MIT (permissive) · 16e5e8f717e9e63b · report fixture: signature model:callable/0/?, model_output_machine_label:str/0/str, pos_bit:int/0/int · returned int · level recorded 2026-09-24
timeit in 59 papers methods/utils.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 58ba64be72bc1b9e · report fixture: signature func:callable/0/? · returned function · level recorded 2026-09-24
tokenize_and_mask in 12 papers methods/abstract_methods/pertubation_based_experiment.py ran MIT (permissive) here; the stored copy is pointer only · 639b00cb6787d6d0 · report fixture: signature text:str/0/str, span_length:int/0/int, buffer_size:int/0/int, pct:float/0/float, ceil_pct:bool/0/bool · returned str · level recorded 2026-09-24
count_masks in 12 papers methods/abstract_methods/pertubation_based_experiment.py unverified MIT (permissive) · 23c3416429f69fee · report
get_supervised_model_prediction methods/abstract_methods/supervised_experiment.py unverified MIT (permissive) here; the stored copy is pointer only · ffed340b4b1b0302 · report
load_multiple_from_file lib/dataset_loader.py unverified MIT (permissive) · 5badef7a89b1732c · report
read_dir_to_pandas lib/dataset_loader.py unverified MIT (permissive) · 9db5a280b863548b · report
replace_masks in 10 papers methods/abstract_methods/pertubation_based_experiment.py unverified MIT (permissive) · 3b992344e4b4f98b · report
trustairlab/mgtbench community (archive-listed) 9 samples, 7 ran (2 instrument-failed) · 2 unverified
check_period dataset_loader.py ran MIT (permissive) · ca626759d96b1cf5 · report fixture: signature texts:list/1/str · returned list · level recorded 2026-09-25
get_lls methods/metric_based.py ran MIT (permissive) · a82535fe98a7e04b · report fixture: signature texts:list/1/str, base_model:callable/0/?, base_tokenizer:callable/0/?, DEVICE:str/0/str · returned list · level recorded 2026-09-25
get_rank in 42 papers methods/metric_based.py ran MIT (permissive) · 0c7cd9182dadcefb · report fixture: signature text:str/0/str, base_model:callable/0/?, base_tokenizer:callable/0/?, DEVICE:str/0/str, log:bool/0/bool · returned float · level recorded 2026-09-25
prim_tree in 5 papers methods/IntrinsicDim.py ran; the contract check could not run (RAISES) MIT (permissive) · 595d1d025c7e692d · report fixture: signature adj_matrix:ndarray/2/float, alpha:float/0/float · returned float · level recorded 2026-09-26
process_string in 8 papers methods/IntrinsicDim.py ran · our draft was wrong fingerprinted MIT (permissive) · ad5f00016ee9222a · report fixture: signature sss:str/0/str · returned str · level recorded 2026-09-26
select_train_data methods/utils.py ran MIT (permissive) · c116b3bbce725a3d · report fixture: signature data:dict/0/?, select_num:int/0/int · returned dict · level recorded 2026-09-25
tokenize_and_mask in 12 papers methods/detectgpt.py ran MIT (permissive) · 4fed12f8c019459e · report fixture: signature text:str/0/str, span_length:int/0/int, buffer_size:int/0/int, pct:float/0/float, ceil_pct:bool/0/bool · returned str · level recorded 2026-09-25
get_ll in 6 papers methods/metric_based.py unverified MIT (permissive) · 3c45c72617228b17 · report
get_supervised_model_prediction methods/supervised.py unverified MIT (permissive) · c4c6872281f5e3a2 · report

Tasks

BenchmarkingQuestion AnsweringSentiment AnalysisText ClassificationText Detectiontext-classification

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections