woog/open-text-detector-qwen36-35b-a3b-oct10
Token Classification • Updated • 2
Datasets:
id string | split_file string | dataset string | kind string | label string | ai_chars int32 | text string | regions list | document_label int8 | supervision string | generator string | source string | venue string | year int16 | text_format string | paper_id string | group string | host_id string | seed_id string | source_start int32 | source_end int32 | target_start int32 | target_end int32 | original_split string | protected bool | draw_id string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
arxiv:9c90575ee94172e6/p01/sonnet-insert_one@105 | stage2-epoch0 | edit_claude | llm_edit | mixed | 111 | uses one specific model only, then that model fixes what the application may itself use. Alternatively, a given CPR library would have to be configurable for each kind of support fault-tolerant model used in a set of production applications, which would be expensive to code, validate, and maintain.
For instance, with ... | [
{
"start": 0,
"end": 1217,
"label": 0
},
{
"start": 1217,
"end": 1328,
"label": 1
},
{
"start": 1328,
"end": 2546,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | 9c90575ee94172e6adc762b9890592b879d5e52687f4e084ceecc433272c25f9 | human:9c90575ee94172e6adc762b9890592b879d5e52687f4e084ceecc433272c25f9 | arxiv:9c90575ee94172e6/p01 | null | 105 | 2,651 | 301 | 1,328 | null | null | stage2-epoch0-edit_claude-0 |
arxiv:9b70c6e3c1a678ba/p00/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 287 | In past years, a common idea of DM physics was that it is related to families of particles beyond the Standard Model (SM), such as those arising from supersymmetric theories (e.g. 6 ). However, extensive LHC runs designed to reveal signatures of events producing non-interacting DM have found no clear evidence of such p... | [
{
"start": 0,
"end": 185,
"label": 0
},
{
"start": 185,
"end": 472,
"label": 1
},
{
"start": 472,
"end": 689,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 9b70c6e3c1a678bab4b3de83af47257dea66ff20de0aaeaee12e3729f338eec7 | human:9b70c6e3c1a678bab4b3de83af47257dea66ff20de0aaeaee12e3729f338eec7 | arxiv:9b70c6e3c1a678ba/p00 | null | 0 | 689 | 0 | 689 | null | null | stage2-epoch0-edit_claude-1 |
pes2o:1d7d4fbc5989552b/p02/sonnet-rewrite_two@504 | stage2-epoch0 | edit_claude | llm_edit | mixed | 640 | samples were obtained from patients with brain tumors and/or chronic pharmacoresistant epilepsy who underwent surgical treatment in the Surgery/Epilepsy Surgery Programs at the University of Bonn Medical Center. In all patients, surgical removal of the tumor/lesion tissue was necessary. Informed consent for additional ... | [
{
"start": 0,
"end": 1306,
"label": 0
},
{
"start": 1306,
"end": 1946,
"label": 1
},
{
"start": 1946,
"end": 2233,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | 1d7d4fbc5989552ba6a6ed8831f2a64593782e85ba75d5e8028fc3738488d690 | human:1d7d4fbc5989552ba6a6ed8831f2a64593782e85ba75d5e8028fc3738488d690 | pes2o:1d7d4fbc5989552b/p02 | null | 504 | 2,737 | 1,131 | 2,233 | null | null | stage2-epoch0-edit_claude-2 |
acl-2019-8f4dde5279215a02/q047/sonnet-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 125 | Cross-Dialectal Modeling: Merged Training Data vs Multitask Learning For the cross-dialectal MSA and E GY models, we first experiment with merging the training datasets for both, and train a single model over the merged datasets. This model is a simple baseline for the cross-dialectal models, but imposes hard joint mod... | [
{
"start": 0,
"end": 1050,
"label": 0
},
{
"start": 1050,
"end": 1175,
"label": 1
},
{
"start": 1175,
"end": 1786,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | proceedings | ACL | 2,019 | paired_pdf_v1 | acl-2019-8f4dde5279215a02 | paper:acl-2019-8f4dde5279215a02 | null | null | 0 | 1,786 | 752 | 1,491 | null | null | stage2-epoch0-edit_claude-3 |
pes2o:4c1d13a7cc9f95a7/p02/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 111 | The above-discussed result confirmed that the benefits of parent involvement should not be underrated. Most Saudi participants took full responsibility for their children's education to ensure that learning in this time of crisis was not compromised. Additionally, they had shown significant consideration for their chil... | [
{
"start": 0,
"end": 1217,
"label": 0
},
{
"start": 1217,
"end": 1328,
"label": 1
},
{
"start": 1328,
"end": 1622,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | 4c1d13a7cc9f95a7d8ac28070281f0ea829ce6dab4ef7a151cf3a2a747327b36 | human:4c1d13a7cc9f95a7d8ac28070281f0ea829ce6dab4ef7a151cf3a2a747327b36 | pes2o:4c1d13a7cc9f95a7/p02 | null | 0 | 1,622 | 780 | 1,622 | null | null | stage2-epoch0-edit_claude-4 |
pmc:4608d7026273ccd5/p01/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 194 | The typical clinical features of oral lichen planus are usually sufficient for the diagnosis of this condition. Still, a biopsy for histopathology is recommended to confirm the clinical diagnosis and mainly to exclude epithelial atypia and signs of malignancy.[@bib18]
The classical histopathological findings in oral l... | [
{
"start": 0,
"end": 1163,
"label": 0
},
{
"start": 1163,
"end": 1357,
"label": 1
},
{
"start": 1357,
"end": 1751,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 4608d7026273ccd545ccb425d9eb2de32f13bb90ee11ebb5c88508add90a4bbd | human:4608d7026273ccd545ccb425d9eb2de32f13bb90ee11ebb5c88508add90a4bbd | pmc:4608d7026273ccd5/p01 | null | 0 | 1,751 | 270 | 1,751 | null | null | stage2-epoch0-edit_claude-5 |
fc2c7c47b918d0c2d792a719dfb602ef/q052/haiku-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 357 | p Corollary 4.4. Let the sampling be defined as follows: S = {i} w. p. p i / M ii , for i 2 [n]. Then there exist acceleration parameters and a Lyapunov function ⌥ k such that f (y k ) f (x ⇤ ) ⌥ k and ⇥ ⇤ p P p k 0 E ⌥ k (1 ⌧ ) k ⌥ 0 = 1 O μ/ i M ii ⌥ .
We now present numerical experiments demonstrating SEGA's ef... | [
{
"start": 0,
"end": 260,
"label": 0
},
{
"start": 260,
"end": 617,
"label": 1
},
{
"start": 617,
"end": 2024,
"label": 0
}
] | null | spans | anthropic/claude-haiku-4-5 | proceedings | NeurIPS | 2,018 | paired_pdf_v1 | fc2c7c47b918d0c2d792a719dfb602ef | paper:fc2c7c47b918d0c2d792a719dfb602ef | null | null | 0 | 2,024 | 260 | 1,123 | null | null | stage2-epoch0-edit_claude-6 |
pes2o:752b4c804efacc63/p04/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 261 | Despite this insight, many patients were uncertain of the quantity and depth of information that they would be comfortable with receiving. The following comment by patient 8 depicts this opinion-'I leave it in the hands of the doctors because they have got me this far'. Both patients and professionals placed value on a... | [
{
"start": 0,
"end": 1697,
"label": 0
},
{
"start": 1697,
"end": 1958,
"label": 1
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | 752b4c804efacc6366310379c2b30ea461698d0ff9467205e55238d38fb4f8c6 | human:752b4c804efacc6366310379c2b30ea461698d0ff9467205e55238d38fb4f8c6 | pes2o:752b4c804efacc63/p04 | null | 0 | 1,958 | 563 | 1,958 | null | null | stage2-epoch0-edit_claude-7 |
arxiv:bb295d2e7944e124/p10/sonnet-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 93 | We advance the literature on consumer comprehension of privacy policies by tracking how length and readability have evolved over two decades and across website ranking tiers.
This section covers applied machine learning. It is the final related area of the literature. A final related area of the privacy policy literat... | [
{
"start": 0,
"end": 176,
"label": 0
},
{
"start": 176,
"end": 269,
"label": 1
},
{
"start": 269,
"end": 921,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | bb295d2e7944e124f1531694d91868e91ba77d14c1f9b3f4ae6780861da83681 | human:bb295d2e7944e124f1531694d91868e91ba77d14c1f9b3f4ae6780861da83681 | arxiv:bb295d2e7944e124/p10 | null | 0 | 921 | 176 | 921 | null | null | stage2-epoch0-edit_claude-8 |
a89b71bb5227c75d463dd82a03115738/q024/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 133 | From now on, we will not distinguish between different ways to define and access the consistent reward polytope P R , but simply assume that we have a polynomial time separation oracle for it. All the algorithms we design in this paper only require access to this separation oracle. In Section 5, we will specify how the... | [
{
"start": 0,
"end": 629,
"label": 0
},
{
"start": 629,
"end": 762,
"label": 1
},
{
"start": 762,
"end": 1203,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | proceedings | NeurIPS | 2,018 | paired_pdf_v1 | a89b71bb5227c75d463dd82a03115738 | paper:a89b71bb5227c75d463dd82a03115738 | null | null | 0 | 1,203 | 358 | 892 | null | null | stage2-epoch0-edit_claude-9 |
pmc:26c2d9ab80f503bd/p03/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 184 | Klussmann ET al. [@bib28] reported a statistically significant difference between HPV prevalence in oropharynx tumor (45%), especially in palatine tonsils (58%) against the other sites of the oral cavity (7 to 25%).
Some authors [@bib11], [@bib28] believe that positive HPV palatine tonsil carcinomas constitute an inde... | [
{
"start": 0,
"end": 342,
"label": 0
},
{
"start": 342,
"end": 526,
"label": 1
},
{
"start": 526,
"end": 1029,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 26c2d9ab80f503bd2a9251c671a808bff9ffd76ae6e7e68af93e4e827e2f69ba | human:26c2d9ab80f503bd2a9251c671a808bff9ffd76ae6e7e68af93e4e827e2f69ba | pmc:26c2d9ab80f503bd/p03 | null | 0 | 1,029 | 217 | 691 | null | null | stage2-epoch0-edit_claude-10 |
pmc:41dba040e28445fa/p01/sonnet-rewrite_two@2218 | stage2-epoch0 | edit_claude | llm_edit | mixed | 349 | Consequently, this large, contemporary population-based survey is initiated to analyze the influence of nontraditional lipid profiles on diabetes in large-scale H-type hypertension population of China.
A study based on city community found an independent contribution of per 1 SD increment in TG/HDL-C to prevalent diab... | [
{
"start": 0,
"end": 1018,
"label": 0
},
{
"start": 1018,
"end": 1367,
"label": 1
},
{
"start": 1367,
"end": 2127,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | 41dba040e28445fa8eb610a3d19625f8f12291f17978a984dfe71844d8a7a7c9 | human:41dba040e28445fa8eb610a3d19625f8f12291f17978a984dfe71844d8a7a7c9 | pmc:41dba040e28445fa/p01 | null | 2,218 | 4,345 | 203 | 1,615 | null | null | stage2-epoch0-edit_claude-11 |
icml-2014-8c87eec1deeb3907/q042/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 138 | Concurrent work by (Mairal, 2013) considers a similar algorithm to that described in Section 2.2, but with H ti a scalar constant rather than a matrix. Proposition 6.1 in (Mairal, 2013) shows that in the case that each g i majorizes its respective f i , and subject to some additional smoothness constraints, G t (x) mon... | [
{
"start": 0,
"end": 1033,
"label": 0
},
{
"start": 1033,
"end": 1171,
"label": 1
},
{
"start": 1171,
"end": 1400,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | proceedings | ICML | 2,014 | paired_pdf_v1 | icml-2014-8c87eec1deeb3907 | paper:icml-2014-8c87eec1deeb3907 | null | null | 0 | 1,400 | 721 | 1,171 | null | null | stage2-epoch0-edit_claude-12 |
arxiv:56ea07c811400397/p01/sonnet-insert_one@1164 | stage2-epoch0 | edit_claude | llm_edit | mixed | 140 | the MC simulation by overlaying inelastic $pp$ interactions produced using Pythia 8 with the NNPDF2.3LO set of PDFs [89] and the A3 set of tuned parameters [90]. The overlaid interactions are generated over a range of multiplicities so as to cover the pile-up conditions encountered during data taking. Events are reweig... | [
{
"start": 0,
"end": 162,
"label": 0
},
{
"start": 162,
"end": 302,
"label": 1
},
{
"start": 302,
"end": 2118,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | 56ea07c811400397c7409e1d2625515be65f58d90b1fb06c93d0fe2bbad96dee | human:56ea07c811400397c7409e1d2625515be65f58d90b1fb06c93d0fe2bbad96dee | arxiv:56ea07c811400397/p01 | null | 1,164 | 3,282 | 0 | 668 | null | null | stage2-epoch0-edit_claude-13 |
arxiv:2b1682356ab25884/p00/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 54 | A (not necessarily polyhedral) outer approximation is a set of constraints that is feasible for all of the points in the set one wishes to approximate. A well known example of an outer approximation is an ILP, where the (initial) linear constraints define an outer approximation of the integer points inside. Outer appro... | [
{
"start": 0,
"end": 605,
"label": 0
},
{
"start": 605,
"end": 659,
"label": 1
},
{
"start": 659,
"end": 1771,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 2b1682356ab2588471e79f4e353cef513c9d883ddee4376518b8645b8d3b8dd3 | human:2b1682356ab2588471e79f4e353cef513c9d883ddee4376518b8645b8d3b8dd3 | arxiv:2b1682356ab25884/p00 | null | 0 | 1,771 | 0 | 975 | null | null | stage2-epoch0-edit_claude-14 |
icml-2020-26612529cee8f4b0/q050/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 110 | N 1 X h l(x n , y n ) + λ M Ω M (x n ) N n=1
Finally, we test the benefits of offline adversarial training (Tramèr et al., 2017a) in addition to the diversity inducing penalties. This allows us to assess whether the two kinds of defense are complementary or whether one subsumes the other. Adversarial examples are pre-... | [
{
"start": 0,
"end": 180,
"label": 0
},
{
"start": 180,
"end": 290,
"label": 1
},
{
"start": 290,
"end": 1199,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | proceedings | ICML | 2,020 | paired_pdf_v1 | icml-2020-26612529cee8f4b0 | paper:icml-2020-26612529cee8f4b0 | null | null | 0 | 1,199 | 46 | 588 | null | null | stage2-epoch0-edit_claude-15 |
pes2o:f05591b06ac1b5fb/p00/sonnet-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 220 | The mammalian defensins comprise the ␣-, -, and -defensin families of cationic, Cys-rich antimicrobial peptides, and each subfamily is characterized by a distinctive tridisulfide array (1). ␣-Defensins are cationic, amphipathic, 3-4-kDa peptides with a -sheet polypeptide backbone and broad spectrum antimicrobial acti... | [
{
"start": 0,
"end": 332,
"label": 0
},
{
"start": 332,
"end": 552,
"label": 1
},
{
"start": 552,
"end": 1909,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | f05591b06ac1b5fb0b8b70d4e3b040cbf36a705aa10fb84770651343d1c78ec4 | human:f05591b06ac1b5fb0b8b70d4e3b040cbf36a705aa10fb84770651343d1c78ec4 | pes2o:f05591b06ac1b5fb/p00 | null | 0 | 1,909 | 0 | 713 | null | null | stage2-epoch0-edit_claude-16 |
arxiv:ad3cc00dc160e261/p00/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 148 | We claim that the increasing sorting order leads to an even distribution of the targets to trajectories. Consequently, the reduction of target coverage after an optimal attack is smaller in the case of OrG-I variants as opposed to the OrG-D ones. We empirically verify the correctness of the above claim by conducting an... | [
{
"start": 0,
"end": 448,
"label": 0
},
{
"start": 448,
"end": 596,
"label": 1
},
{
"start": 596,
"end": 843,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | ad3cc00dc160e261dce4faf38a54fb240c17cff57294939eca278e1c37d63bbb | human:ad3cc00dc160e261dce4faf38a54fb240c17cff57294939eca278e1c37d63bbb | arxiv:ad3cc00dc160e261/p00 | null | 0 | 843 | 0 | 843 | null | null | stage2-epoch0-edit_claude-17 |
pmc:0dc911c7dd4c9ee9/p00/opus-rewrite_two@538 | stage2-epoch0 | edit_claude | llm_edit | mixed | 289 | ended up not being empowering but a restricting gap between law and the practice of governance? This concern has also been raised by Scott and Holder ([@CR95]). Furthermore, did this flexibility fail to facilitate the adoption of the systems thinking needed for the participatory ecological design expected by the WFD pr... | [
{
"start": 0,
"end": 96,
"label": 0
},
{
"start": 96,
"end": 385,
"label": 1
},
{
"start": 385,
"end": 2578,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 0dc911c7dd4c9ee992d376646dafaaa94c5c5fe41605f67680ba712cb624acce | human:0dc911c7dd4c9ee992d376646dafaaa94c5c5fe41605f67680ba712cb624acce | pmc:0dc911c7dd4c9ee9/p00 | null | 538 | 3,116 | 0 | 530 | null | null | stage2-epoch0-edit_claude-18 |
pmc:f218a98bdf0709c8/p02/opus-insert_one@1841 | stage2-epoch0 | edit_claude | llm_edit | mixed | 126 | CI, 2.90 to 8.68) greater than the Tx group. The results demonstrated a better therapeutic effect in FEV1% predicted with clinical relevance of Mx + Tx and A + Tx compared to Tx.
Wang et al. suggested that acupuncture therapy may be effective and safe in improving HRQL and pulmonary function in patients with COPD ([@B... | [
{
"start": 0,
"end": 1393,
"label": 0
},
{
"start": 1393,
"end": 1519,
"label": 1
},
{
"start": 1519,
"end": 2340,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | f218a98bdf0709c85ad4ab3e98247a025067dcfb4888b711a7a60b0193efcef2 | human:f218a98bdf0709c85ad4ab3e98247a025067dcfb4888b711a7a60b0193efcef2 | pmc:f218a98bdf0709c8/p02 | null | 1,841 | 4,181 | 180 | 1,571 | null | null | stage2-epoch0-edit_claude-19 |
pmc:50fa88f37cdb9a12/p01/sonnet-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 122 | For this analysis, due to the distribution and non-linear response, associations were examined using approximate percentile cut-offs rather than the continuous cognitive score. Poor performance was defined as obtaining a score less than a cut-off point corresponding to approximately the 25th percentile of the populatio... | [
{
"start": 0,
"end": 985,
"label": 0
},
{
"start": 985,
"end": 1107,
"label": 1
},
{
"start": 1107,
"end": 1776,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | 50fa88f37cdb9a12c2ea7cd63107c781de0c40c83cb61d34b64a28c757374b46 | human:50fa88f37cdb9a12c2ea7cd63107c781de0c40c83cb61d34b64a28c757374b46 | pmc:50fa88f37cdb9a12/p01 | null | 0 | 1,776 | 566 | 1,776 | null | null | stage2-epoch0-edit_claude-20 |
pes2o:cd47d37477080271/p01/haiku-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 338 | Physical interpretation of shear-rate behaviour of soils and geotechnical solution to the coefficient of start-up friction with low inertial number
Shear experiments on soils have revealed the effects of shear rate, confining pressure, and grain size on the residual shear strength, but their nature is not well underst... | [
{
"start": 0,
"end": 953,
"label": 0
},
{
"start": 953,
"end": 1291,
"label": 1
},
{
"start": 1291,
"end": 1412,
"label": 0
}
] | null | spans | anthropic/claude-haiku-4-5 | pes2o | null | null | host_pes2o | cd47d374770802717ecf85272101bc40d889d8d059c1c93c69a457d703d04418 | human:cd47d374770802717ecf85272101bc40d889d8d059c1c93c69a457d703d04418 | pes2o:cd47d37477080271/p01 | null | 0 | 1,412 | 149 | 1,412 | null | null | stage2-epoch0-edit_claude-21 |
arxiv:482e5301de5b737c/p02/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 110 | This presents a trade-off to the maintainers of Linux distributions. On the one hand, adopting new versions of packages within the distribution will grant users access to new features, bug fixes and security patches. On the other hand, these new versions risk introducing breaking changes, new bugs, security vulnerabili... | [
{
"start": 0,
"end": 1622,
"label": 0
},
{
"start": 1622,
"end": 1732,
"label": 1
},
{
"start": 1732,
"end": 1897,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | 482e5301de5b737c16569db79d157d321d3db5f3b9b7fa6377be81a362984b7f | human:482e5301de5b737c16569db79d157d321d3db5f3b9b7fa6377be81a362984b7f | arxiv:482e5301de5b737c/p02 | null | 0 | 1,897 | 779 | 1,897 | null | null | stage2-epoch0-edit_claude-22 |
arxiv:8e3bf1235737a9b8/p00/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 342 | Because most e-commerce companies handle a large volume of searches per second, performing part of the model training offline is important for efficiency. Representation-based models can compute document embeddings independently of queries (offline), which gives them an important advantage over interaction-based models... | [
{
"start": 0,
"end": 342,
"label": 1
},
{
"start": 342,
"end": 917,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 8e3bf1235737a9b8b3e65226daa9b17667a2e64ecc1b947423d404a6def711b0 | human:8e3bf1235737a9b8b3e65226daa9b17667a2e64ecc1b947423d404a6def711b0 | arxiv:8e3bf1235737a9b8/p00 | null | 0 | 917 | 0 | 475 | null | null | stage2-epoch0-edit_claude-23 |
acl-2020-a5f099e8a63298c9/q011/sonnet-rewrite_two@0 | stage2-epoch0 | edit_claude | llm_edit | mixed | 141 | than a derivation structure, makes it possible to define syntactic parsing as a sequenceto-sequence problem, mapping the sentence to its parse sequence. If a neural network architecture for modelling sequences (called seq2seq models) can perform well at this task, then maybe the structured linguistic representations of... | [
{
"start": 0,
"end": 1187,
"label": 0
},
{
"start": 1187,
"end": 1328,
"label": 1
},
{
"start": 1328,
"end": 2580,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | proceedings | ACL | 2,020 | paired_pdf_v1 | acl-2020-a5f099e8a63298c9 | paper:acl-2020-a5f099e8a63298c9 | null | null | 0 | 2,580 | 900 | 2,580 | null | null | stage2-epoch0-edit_claude-24 |
or:0f4dd6cba2340e11/p015/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 176 | The work of Koppel et al. (2005) is one of the first ethno-nationality identification work where it involves categorizing users based on their native language employing Support Vector Machines (SVMs) on various stylistic features focused on identification of common. Authors achieved 0.802 accuracy using five chosen lan... | [
{
"start": 0,
"end": 752,
"label": 0
},
{
"start": 752,
"end": 928,
"label": 1
},
{
"start": 928,
"end": 1612,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | openreview | ICLR | 2,022 | clean_v2 | 0f4dd6cba2340e112130b4ead22b4d2819edf9db531e91f5db84400bdb446b18 | human:0f4dd6cba2340e112130b4ead22b4d2819edf9db531e91f5db84400bdb446b18 | or:0f4dd6cba2340e11/p015 | null | 0 | 1,612 | 592 | 1,016 | null | null | stage2-epoch0-edit_claude-25 |
pes2o:786f43ed4f191afc/p03/sonnet-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 150 | Media
Media used for antibacterial studies was Mueller-Hinton's [15] agar and the technique used was agar well diffusion [16] .The stock solution of methanol and diethyl ether extract made from seed powder and callus of Pongamia pinnata was used. The extracts were further diluted so as to get concentrations in the ran... | [
{
"start": 0,
"end": 706,
"label": 0
},
{
"start": 706,
"end": 856,
"label": 1
},
{
"start": 856,
"end": 1463,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | 786f43ed4f191afc0e16eea13f068636c041eb9d59b2b4a35272ae56d413e9c9 | human:786f43ed4f191afc0e16eea13f068636c041eb9d59b2b4a35272ae56d413e9c9 | pes2o:786f43ed4f191afc/p03 | null | 0 | 1,463 | 7 | 1,440 | null | null | stage2-epoch0-edit_claude-26 |
pes2o:f3afb462759e202a/p01/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 92 | C. Inceptionresnetv2 OF DeIN ARCHITECTURE
In inceptionResnetV2, deformable convolution simply expressed as Figure 6(a), meanwhile, inception-ResnetA and inception-ResnetB which include max-pooling layers, but in DeIN, these max-pooling layers become deformable max-pooling layer by adding offset layer, which increased ... | [
{
"start": 0,
"end": 714,
"label": 0
},
{
"start": 714,
"end": 806,
"label": 1
},
{
"start": 806,
"end": 1514,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | f3afb462759e202a3709b10ccfa30bd0ed882ae19e11b7a64d56338d139c698f | human:f3afb462759e202a3709b10ccfa30bd0ed882ae19e11b7a64d56338d139c698f | pes2o:f3afb462759e202a/p01 | null | 0 | 1,514 | 43 | 891 | null | null | stage2-epoch0-edit_claude-27 |
acl-2019-cf09e85f11bb4a91/q018/sonnet-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 65 | prepend a wh* word (randomly or with the heuristic in Section 2.4) and append a question mark. The noise function consists of word dropout, word order permutation and word masking. The motivation is that, at least for SQuAD, it may be sufficient to simply learn a function to identify a span surrounded by high n-gram ov... | [
{
"start": 0,
"end": 566,
"label": 0
},
{
"start": 566,
"end": 631,
"label": 1
},
{
"start": 631,
"end": 1194,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | proceedings | ACL | 2,019 | paired_pdf_v1 | acl-2019-cf09e85f11bb4a91 | paper:acl-2019-cf09e85f11bb4a91 | null | null | 0 | 1,194 | 390 | 822 | null | null | stage2-epoch0-edit_claude-28 |
pmc:e99c23b12e093f3d/p01/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 350 | The 14 districts in the second spatial-temporal cluster, which are mainly located in the areas of cold spots from 2012 to 2016, had consistently low dengue incidences between June 2013 and November 2015. These 14 districts are mainly located in the mid-north and south of Sri Lanka. The humidity and wind speed were foun... | [
{
"start": 0,
"end": 928,
"label": 0
},
{
"start": 928,
"end": 1278,
"label": 1
},
{
"start": 1278,
"end": 1923,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | e99c23b12e093f3d4b94130810c3d8f5483c5feedf0bb09ea853c69ef99640b6 | human:e99c23b12e093f3d4b94130810c3d8f5483c5feedf0bb09ea853c69ef99640b6 | pmc:e99c23b12e093f3d/p01 | null | 0 | 1,923 | 540 | 1,498 | null | null | stage2-epoch0-edit_claude-29 |
or:b7720f85c7e3780e/p003/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 138 | ABSTRACT
Learning image representations on decentralized data can bring many benefits in cases where data cannot be aggregated across data silos. Softmax cross entropy loss is highly effective and commonly used for learning image representations. Using a large number of classes has proven to be particularly beneficial... | [
{
"start": 0,
"end": 1182,
"label": 0
},
{
"start": 1182,
"end": 1320,
"label": 1
},
{
"start": 1320,
"end": 1323,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | openreview | ICLR | 2,022 | clean_v2 | b7720f85c7e3780e6e3d740555841722c5d1669835e4e28d7c7c83452bc4ac19 | human:b7720f85c7e3780e6e3d740555841722c5d1669835e4e28d7c7c83452bc4ac19 | or:b7720f85c7e3780e/p003 | null | 0 | 1,323 | 10 | 1,320 | null | null | stage2-epoch0-edit_claude-30 |
pmc:89f0d12093d0047b/p00/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 75 | Human dermal microvascular endothelial cells (HMECs) obtained from the Centers for Disease Control and Prevention (Atlanta, GA) were grown in MCDB 131 medium (Caisson\'s Laboratories) containing 10% fetal bovine serum (Aleken Biologicals), 10 ng/ml epidermal growth factor (Thermo Fisher Scientific), 1 *μ*g/ml hydrocort... | [
{
"start": 0,
"end": 907,
"label": 0
},
{
"start": 907,
"end": 982,
"label": 1
},
{
"start": 982,
"end": 1419,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | 89f0d12093d0047b02a504d83e857370c78ec634ad7841438c9eff55078c2f00 | human:89f0d12093d0047b02a504d83e857370c78ec634ad7841438c9eff55078c2f00 | pmc:89f0d12093d0047b/p00 | null | 0 | 1,419 | 0 | 1,419 | null | null | stage2-epoch0-edit_claude-31 |
arxiv:8e33e9d0296f9a2b/p01/opus-rewrite_one@0 | stage2-epoch0 | edit_claude | llm_edit | mixed | 271 | A network of operating and designed gravitational wave detectors raise expectations that, sooner or later, the relic gravitational radiation background will be detected. Gravitational waves (GWs) are thought to be generated by several processes in the early universe Kuroyanagi _et al._ (2018); Christensen (2019); Capri... | [
{
"start": 0,
"end": 1092,
"label": 0
},
{
"start": 1092,
"end": 1363,
"label": 1
},
{
"start": 1363,
"end": 1909,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 8e33e9d0296f9a2b0d10a5e12ff66a32b8ea24232c960d1226ac095271dddd16 | human:8e33e9d0296f9a2b0d10a5e12ff66a32b8ea24232c960d1226ac095271dddd16 | arxiv:8e33e9d0296f9a2b/p01 | null | 0 | 1,909 | 664 | 1,363 | null | null | stage2-epoch0-edit_claude-32 |
arxiv:95f2251710d91cec/p00/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 213 | More related to our paper, Allcott et al., (2020) model differences in beliefs and attitudes as resulting from messages on the crisis from both political leaders and the media. They use survey data and show that Republicans and Democrats engage in social distancing to different extents. Their theoretical model is also ... | [
{
"start": 0,
"end": 524,
"label": 0
},
{
"start": 524,
"end": 737,
"label": 1
},
{
"start": 737,
"end": 1708,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 95f2251710d91cecd94dc0bb57a068ba0c6280e17c60ffa32fba0e01210e436c | human:95f2251710d91cecd94dc0bb57a068ba0c6280e17c60ffa32fba0e01210e436c | arxiv:95f2251710d91cec/p00 | null | 0 | 1,708 | 0 | 737 | null | null | stage2-epoch0-edit_claude-33 |
arxiv:ce22188d13349690/p03/opus-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 156 | The width and lifetime of the strands that we see in the synthesized observations is compatible with observed threads in coronal loops (Brooks et al., 2013; Williams et al., 2020b, a). Strand widths range from 250 to 1400 km FWHM for the AIA 211 Å channel and 600 to 1500 km FWHM for the XRT channel. Brooks et al. (2013... | [
{
"start": 0,
"end": 942,
"label": 0
},
{
"start": 942,
"end": 1098,
"label": 1
},
{
"start": 1098,
"end": 1563,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | ce22188d13349690017f10f66a2f0aa0741e626538888c0415972b14290e1511 | human:ce22188d13349690017f10f66a2f0aa0741e626538888c0415972b14290e1511 | arxiv:ce22188d13349690/p03 | null | 0 | 1,563 | 942 | 1,563 | null | null | stage2-epoch0-edit_claude-34 |
arxiv:9cc94701224e5f9e/p00/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 106 | We use the publicly available OpenI dataset444https://openi.nlm.nih.gov/services#searchAPIUsingGET [1] collected by Indiana Network for Patient Care. It comprises 3,995 annotated radiology reports, together with 7,470 radiology images associated with them. To the best of our knowledge, there are no large corpora of ann... | [
{
"start": 0,
"end": 150,
"label": 0
},
{
"start": 150,
"end": 256,
"label": 1
},
{
"start": 256,
"end": 1161,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 9cc94701224e5f9e76d263fc75a12e821453d38bdea0da4f59c1f58516f0c9b4 | human:9cc94701224e5f9e76d263fc75a12e821453d38bdea0da4f59c1f58516f0c9b4 | arxiv:9cc94701224e5f9e/p00 | null | 0 | 1,161 | 0 | 1,161 | null | null | stage2-epoch0-edit_claude-35 |
or:10161229403f93db/p145/sonnet-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 98 | Q πh,β (s, a) − Q ∗ h,β (s, a) ≤ 2(T − h + 1)(δ r + ε), h ≤ T.
These results are proved in Appendix B.2 for the episodic setting and in Appendix B.3 for the discounted setting. Note also that these bounds are tight (see, e.g. Nemecek and Parr [31] for the risk-neutral case). Finally, while not required in this work, t... | [
{
"start": 0,
"end": 178,
"label": 0
},
{
"start": 178,
"end": 276,
"label": 1
},
{
"start": 276,
"end": 495,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | NeurIPS | 2,021 | clean_v2 | 10161229403f93db7a3b7091184687fd40285b902199382ca1ddc847f6fe1886 | human:10161229403f93db7a3b7091184687fd40285b902199382ca1ddc847f6fe1886 | or:10161229403f93db/p145 | null | 0 | 495 | 64 | 490 | null | null | stage2-epoch0-edit_claude-36 |
pmc:09868c36d2f7f586/p02/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 221 | The standard diagnosis of NASH is made by liver biopsy and histopathological staging \[[@CR9], [@CR10]\]. Currently different imaging methods are used for evaluation and quantification of NAFLD, as biopsy is invasive with risk of complications, semi-quantitative and prone to sample variability. Fibroscan™ as a method f... | [
{
"start": 0,
"end": 1313,
"label": 0
},
{
"start": 1313,
"end": 1534,
"label": 1
},
{
"start": 1534,
"end": 1743,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 09868c36d2f7f586d011881388ced2ddeb6677bb79bb9a4cebcb5416a39d8d9d | human:09868c36d2f7f586d011881388ced2ddeb6677bb79bb9a4cebcb5416a39d8d9d | pmc:09868c36d2f7f586/p02 | null | 0 | 1,743 | 872 | 1,743 | null | null | stage2-epoch0-edit_claude-37 |
or:df77912eb35cf659/p009/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 117 | Extensive empirical analysis of pruning techniques has inspired associated theoretical developments. Several works have derived bounds for the performance and size of subnetworks discovered in randomly-initialized networks (Malach et al., 2020; Orseau et al., 2020; Pensia et al., 2020). Other theoretical works analyze ... | [
{
"start": 0,
"end": 1111,
"label": 0
},
{
"start": 1111,
"end": 1228,
"label": 1
},
{
"start": 1228,
"end": 2013,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | ICLR | 2,022 | clean_v2 | df77912eb35cf659c48ecd33389bee95cc56ce047ec52f35659ef3164748683a | human:df77912eb35cf659c48ecd33389bee95cc56ce047ec52f35659ef3164748683a | or:df77912eb35cf659/p009 | null | 0 | 2,013 | 920 | 2,010 | null | null | stage2-epoch0-edit_claude-38 |
pmc:1e6f316bdaf44b7f/p02/opus-rewrite_one@0 | stage2-epoch0 | edit_claude | llm_edit | mixed | 143 | HFA evaluates pure tone air conduction thresholds in the 9--20 kHz range, depending on the equipment. This is an important test for the early detection of hearing loss caused by damage to the base of the cochlea, as occurs in cases of ototoxicity, because it allows the detection of hearing loss before reaching the conv... | [
{
"start": 0,
"end": 1121,
"label": 0
},
{
"start": 1121,
"end": 1264,
"label": 1
},
{
"start": 1264,
"end": 2121,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 1e6f316bdaf44b7fd2dff3212fe4784e6a76d6fad3b45a35cd0264697ef4f759 | human:1e6f316bdaf44b7fd2dff3212fe4784e6a76d6fad3b45a35cd0264697ef4f759 | pmc:1e6f316bdaf44b7f/p02 | null | 0 | 2,121 | 395 | 1,721 | null | null | stage2-epoch0-edit_claude-39 |
pes2o:f74f3d95ac3d55e2/p01/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 540 | Enumeration of CTC by CellSearch
Circulating tumour cell enumeration using the CellSearch system (Veridex LLC, Raritan, NJ, USA) was carried out according to the manufacturer's protocol and training. Blood samples were processed on the CellTracks Autoprep within 72 h. Epithelial cells were immunomagnetically enriched ... | [
{
"start": 0,
"end": 951,
"label": 0
},
{
"start": 951,
"end": 1491,
"label": 1
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | f74f3d95ac3d55e2f2673abb3dd572c14fd0d226ce4dcdbe83641e71281fa1e6 | human:f74f3d95ac3d55e2f2673abb3dd572c14fd0d226ce4dcdbe83641e71281fa1e6 | pes2o:f74f3d95ac3d55e2/p01 | null | 0 | 1,491 | 34 | 1,491 | null | null | stage2-epoch0-edit_claude-40 |
pes2o:9b53e852d6ca0baa/p04/sonnet-insert_one@2292 | stage2-epoch0 | edit_claude | llm_edit | mixed | 80 | www.nature.com/scientificreports/ defining the functional state of cells and tissues, enable the validation and assessment of resistance mechanisms in clinical samples 5,54 . High attrition rates in clinical drug development present significant challenges to drug developers, with only one in eight oncology drugs that e... | [
{
"start": 0,
"end": 2394,
"label": 0
},
{
"start": 2394,
"end": 2474,
"label": 1
},
{
"start": 2474,
"end": 2500,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | 9b53e852d6ca0baa46696fb439bc1c4334fa2ff0ce374cbcd3b1f911861b5b30 | human:9b53e852d6ca0baa46696fb439bc1c4334fa2ff0ce374cbcd3b1f911861b5b30 | pes2o:9b53e852d6ca0baa/p04 | null | 2,292 | 4,792 | 1,295 | 2,500 | null | null | stage2-epoch0-edit_claude-41 |
or:61c7aabe87edfe75/p008/sonnet-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 228 | ABSTRACT
We characterize the power-law asymptotics of learning curves for Gaussian process regression (GPR), assuming that both the eigenspectrum of the prior and the eigenexpansion coefficients of the target function follow a power law. Under similar assumptions, we leverage the equivalence between GPR and kernel rid... | [
{
"start": 0,
"end": 10,
"label": 0
},
{
"start": 10,
"end": 238,
"label": 1
},
{
"start": 238,
"end": 731,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | ICLR | 2,022 | clean_v2 | 61c7aabe87edfe75e0b15b4fba3f4251b5c11bd5f4944851413bef7bb1ff83de | human:61c7aabe87edfe75e0b15b4fba3f4251b5c11bd5f4944851413bef7bb1ff83de | or:61c7aabe87edfe75/p008 | null | 0 | 731 | 10 | 728 | null | null | stage2-epoch0-edit_claude-42 |
pes2o:f7dd4396374a8fdd/p11/haiku-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 186 | As successfully defending against vanilla attacks does not really guarantee privacy, we considered adaptive attacks when evaluating all defenses, which means the attackers have the knowledge of which defense algorithm and what hyper-parameters may have been applied to training the target model, so that they can plug th... | [
{
"start": 0,
"end": 2071,
"label": 0
},
{
"start": 2071,
"end": 2257,
"label": 1
}
] | null | spans | anthropic/claude-haiku-4-5 | pes2o | null | null | host_pes2o | f7dd4396374a8fddb6daf15039641c7d766e87a477f006b6d71f3378495df87f | human:f7dd4396374a8fddb6daf15039641c7d766e87a477f006b6d71f3378495df87f | pes2o:f7dd4396374a8fdd/p11 | null | 0 | 2,257 | 1,468 | 2,257 | null | null | stage2-epoch0-edit_claude-43 |
arxiv:ab84a2065386dbcf/p00/sonnet-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 302 | The aim of this paper is to characterize the immersion of a surface in a given simply connected complete Lorentzian homogeneous space of dimension $3$ in terms of spinors. Here we focus on Lorentzian metric Lie groups and on symmetric spaces that have constant curvature or are products, leaving the Cahen-Wallach spaces... | [
{
"start": 0,
"end": 172,
"label": 0
},
{
"start": 172,
"end": 474,
"label": 1
},
{
"start": 474,
"end": 1188,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | ab84a2065386dbcf1d0b84dae3476d0bb2a8de8d31bebe82a0f1362b0ac22e88 | human:ab84a2065386dbcf1d0b84dae3476d0bb2a8de8d31bebe82a0f1362b0ac22e88 | arxiv:ab84a2065386dbcf/p00 | null | 0 | 1,188 | 0 | 474 | null | null | stage2-epoch0-edit_claude-44 |
pmc:88fd0e54aff4f7b6/p00/haiku-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 121 | Our work focused on the social conditions that are associated with uneven exposure to SARS-CoV-2, and higher infections, for disadvantaged communities in the spring of 2020. Since then, this situation has been compounded by national interventions that do not appear to be yielding equitable outcomes. For example, the fe... | [
{
"start": 0,
"end": 824,
"label": 0
},
{
"start": 824,
"end": 945,
"label": 1
},
{
"start": 945,
"end": 1356,
"label": 0
}
] | null | spans | anthropic/claude-haiku-4-5 | pmc | null | null | host_pmc | 88fd0e54aff4f7b6b7fe70b31221077ea8f613c7d70f205bc6471fd85eb40b58 | human:88fd0e54aff4f7b6b7fe70b31221077ea8f613c7d70f205bc6471fd85eb40b58 | pmc:88fd0e54aff4f7b6/p00 | null | 0 | 1,356 | 0 | 1,356 | null | null | stage2-epoch0-edit_claude-45 |
pmc:5c348fa7341c816b/p00/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 212 | Several habitation structures at the site provided a small sample of faunal remains that shed light on the earliest history of animal domesticate use in the Alay Valley. Layer 3, the lowest stratigraphic unit, consists of a yellowish loess deposit. This layer is directly dated by charcoal recovered from the base of the... | [
{
"start": 0,
"end": 170,
"label": 0
},
{
"start": 170,
"end": 382,
"label": 1
},
{
"start": 382,
"end": 1950,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 5c348fa7341c816bddec00c4ebb8ba7f6ef1864ff8766fe2e06d4e449021648d | human:5c348fa7341c816bddec00c4ebb8ba7f6ef1864ff8766fe2e06d4e449021648d | pmc:5c348fa7341c816b/p00 | null | 0 | 1,950 | 0 | 850 | null | null | stage2-epoch0-edit_claude-46 |
pes2o:064b7b2510c7914d/p01/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 109 | THE PHENOLOGICAL AND POMOLOGICAL TRAITS OF SELECTED GENOTYPES OF WILD PEAR [ Pyrus pyraster ( L . ) Du Roi ] IMPORTANT FOR THE PRODUCTION OF GENERATIVE ROOTSTOCKS
The aim of this work is to produce the generative rootstocks for the wild pears, which will be of moderate denseness. The initial material is the population... | [
{
"start": 0,
"end": 1622,
"label": 0
},
{
"start": 1622,
"end": 1731,
"label": 1
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | 064b7b2510c7914d7248a7ab2937cfc650a3df0f09c6eb240a8a343bc37125d4 | human:064b7b2510c7914d7248a7ab2937cfc650a3df0f09c6eb240a8a343bc37125d4 | pes2o:064b7b2510c7914d/p01 | null | 0 | 1,731 | 164 | 1,731 | null | null | stage2-epoch0-edit_claude-47 |
pmc:d08cd5743bb6719c/p00/opus-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 170 | Left atrial volume is significantly greater in endurance athletes compared to strength-trained athletes and the normal population \[[@CR30]--[@CR32]\], and has been identified as a predictor of AF in athletes \[[@CR11], [@CR33]\]. Pellicia and colleagues demonstrated in a cohort of 1777 athletes from a mixture of sport... | [
{
"start": 0,
"end": 628,
"label": 0
},
{
"start": 628,
"end": 798,
"label": 1
},
{
"start": 798,
"end": 1370,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | d08cd5743bb6719cf509318212b429cd3964b81df887e109a3685b4c1f9646a6 | human:d08cd5743bb6719cf509318212b429cd3964b81df887e109a3685b4c1f9646a6 | pmc:d08cd5743bb6719c/p00 | null | 0 | 1,370 | 0 | 1,370 | null | null | stage2-epoch0-edit_claude-48 |
arxiv:5eabcff95e311b56/p00/opus-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 244 | When the goal is to approximate the “true” indices, Janon/Monod, Jansen and Azzini and Rosati also offer the best performance. The median MAE produced by these estimators is generally smaller than that of Glen and Isaacs’ and pseudo- Owen’s. Their MAE values are also far more tightly distributed than those obtained wit... | [
{
"start": 0,
"end": 127,
"label": 0
},
{
"start": 127,
"end": 371,
"label": 1
},
{
"start": 371,
"end": 1736,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 5eabcff95e311b56bd932db5da14be19ee280cec5e2c99c1249495af715b54a0 | human:5eabcff95e311b56bd932db5da14be19ee280cec5e2c99c1249495af715b54a0 | arxiv:5eabcff95e311b56/p00 | null | 0 | 1,736 | 0 | 698 | null | null | stage2-epoch0-edit_claude-49 |
pes2o:59f6688b2481a0f6/p01/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 103 | Introduction
Congenital heart defects (CHDs) are defects in the structure of the heart and great vessels that occur at birth. Errors in septation, proper patterning of the great vessels, and valve formation are the most common aberrations in cardiac development that lead to the majority of CHDs seen. This progress mea... | [
{
"start": 0,
"end": 303,
"label": 0
},
{
"start": 303,
"end": 406,
"label": 1
},
{
"start": 406,
"end": 1618,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | 59f6688b2481a0f6389b0837b9c456ddadd9719d9323d2707cf0256059761df4 | human:59f6688b2481a0f6389b0837b9c456ddadd9719d9323d2707cf0256059761df4 | pes2o:59f6688b2481a0f6/p01 | null | 0 | 1,618 | 14 | 730 | null | null | stage2-epoch0-edit_claude-50 |
pmc:d0242ba1d583b1ac/p01/sonnet-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 270 | According to Kemp,[@bib28] TOAE responses are stronger and more easily detected in the 1 to 4 kHz range; a normal adult ear yields weak TOAE (less than 3 dBSPL) and no substantial responses above 4 kHz. Mor and Azevedo[@bib29] also found a uniformly occurring TOAE response between 1 000 and 4 000 Hz, decreasing at 5 00... | [
{
"start": 0,
"end": 420,
"label": 0
},
{
"start": 420,
"end": 690,
"label": 1
},
{
"start": 690,
"end": 984,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | d0242ba1d583b1ac1be92900d869d9e6f08406c02593373a9ee50b069fbd277c | human:d0242ba1d583b1ac1be92900d869d9e6f08406c02593373a9ee50b069fbd277c | pmc:d0242ba1d583b1ac/p01 | null | 0 | 984 | 327 | 984 | null | null | stage2-epoch0-edit_claude-51 |
pmc:0e72acde4c647c19/p00/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 132 | Vitamin-D plays several roles in the growth of children ([@B14], [@B18]). First, children need higher calcium intake than adults to provide adequate calcium for bone mineralization ([@B35]). Vitamin D contributes to bone mineralization by keeping blood calcium and phosphate concentrations within the normal range ([@B36... | [
{
"start": 0,
"end": 191,
"label": 0
},
{
"start": 191,
"end": 323,
"label": 1
},
{
"start": 323,
"end": 1253,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 0e72acde4c647c19d3488cd399503f6961a44406c965b21c263d8e467571ee6d | human:0e72acde4c647c19d3488cd399503f6961a44406c965b21c263d8e467571ee6d | pmc:0e72acde4c647c19/p00 | null | 0 | 1,253 | 0 | 808 | null | null | stage2-epoch0-edit_claude-52 |
arxiv:e64c85ffe368850f/p02/sonnet-rewrite_one@82 | stage2-epoch0 | edit_claude | llm_edit | mixed | 77 | given _sink_ set $R\subseteq S$. Intuitively, end components are parts of the state space where the system can remain forever under a particular strategy. For example, suppose there are two states $s_{1},s_{2}$ where $\Delta(s_{1},a_{1},s_{2})=\Delta(s_{2},a_{2},s_{1})=1$, i.e. the system can go back and forth between ... | [
{
"start": 0,
"end": 1651,
"label": 0
},
{
"start": 1651,
"end": 1728,
"label": 1
},
{
"start": 1728,
"end": 1938,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | e64c85ffe368850fa648ddcd0308076f7fc4980feb1c0e5befee590eaaead93b | human:e64c85ffe368850fa648ddcd0308076f7fc4980feb1c0e5befee590eaaead93b | arxiv:e64c85ffe368850f/p02 | null | 82 | 2,020 | 1,312 | 1,938 | null | null | stage2-epoch0-edit_claude-53 |
arxiv:01217ebd848ffca7/p01/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 110 | The results are presented in the “Rank” column of Table 2, where the highest rank corresponds to the most adopted practice. We observe that the most adopted practices (practices 6, 7) are related to establishing and communicating clear objectives and metrics for training. Versioning (practice 16), continuous monitoring... | [
{
"start": 0,
"end": 594,
"label": 0
},
{
"start": 594,
"end": 704,
"label": 1
},
{
"start": 704,
"end": 1202,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 01217ebd848ffca7b2f9a142b282fd4dc699249eba7a7e748ab68d785845ae23 | human:01217ebd848ffca7b2f9a142b282fd4dc699249eba7a7e748ab68d785845ae23 | arxiv:01217ebd848ffca7/p01 | null | 0 | 1,202 | 594 | 1,202 | null | null | stage2-epoch0-edit_claude-54 |
pmc:481c70caf05d7e1b/p00/sonnet-insert_one@299 | stage2-epoch0 | edit_claude | llm_edit | mixed | 95 | [@bib0270], [@bib0275], [@bib0280], [@bib0285] with differences in the studied populations and in data collection format. Interestingly, all of the complications reported here occurred in the first two years of flap use and in adult patients who had the third molars. Some groups[@bib0265], [@bib0275] have performed mod... | [
{
"start": 0,
"end": 544,
"label": 0
},
{
"start": 544,
"end": 639,
"label": 1
},
{
"start": 639,
"end": 2039,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | 481c70caf05d7e1bd385d87a50c4a0debacf66bf86291c0d663a07b9a085a320 | human:481c70caf05d7e1bd385d87a50c4a0debacf66bf86291c0d663a07b9a085a320 | pmc:481c70caf05d7e1b/p00 | null | 299 | 2,338 | 0 | 1,321 | null | null | stage2-epoch0-edit_claude-55 |
pmc:cc06ae20a3156bae/p03/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 107 | We must stress that, during the 10 recordings carried out in each animal of the control-group, we observed a constant reproducibility in the findings, showing that the results are reliable and, moreover, they may help establish a standard concept of normality for this study in bovines.
Animals currently used for resea... | [
{
"start": 0,
"end": 513,
"label": 0
},
{
"start": 513,
"end": 620,
"label": 1
},
{
"start": 620,
"end": 1781,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | cc06ae20a3156bae70039b0b880cea2322aaa04fb544791bb2ca4ad68c137792 | human:cc06ae20a3156bae70039b0b880cea2322aaa04fb544791bb2ca4ad68c137792 | pmc:cc06ae20a3156bae/p03 | null | 0 | 1,781 | 288 | 1,567 | null | null | stage2-epoch0-edit_claude-56 |
pmc:b6c3d1a675d04766/p02/opus-rewrite_two@790 | stage2-epoch0 | edit_claude | llm_edit | mixed | 140 | in the system become similar. A similarity in the phase diagrams was also observed between PEG-400 at a weight fraction of 0.2 and PEG-1000 at a weight fraction of 0.08 in the system. The results indicate that the phase diagram of the ternary system of gelatin, water and PEG can be controlled by the appropriate choice ... | [
{
"start": 0,
"end": 1178,
"label": 0
},
{
"start": 1178,
"end": 1318,
"label": 1
},
{
"start": 1318,
"end": 1927,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | b6c3d1a675d0476676b5c7d58de308ec67780cf76393671b2b4bca90aa9db2c5 | human:b6c3d1a675d0476676b5c7d58de308ec67780cf76393671b2b4bca90aa9db2c5 | pmc:b6c3d1a675d04766/p02 | null | 790 | 2,717 | 386 | 1,927 | null | null | stage2-epoch0-edit_claude-57 |
or:9e0f95a67c44371e/p092/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 120 | The goal here is to boost the predictions of a pre-trained classifier through multiple rounds of auditing (searching for worst-case subgroups using an auxiliary model) rather than learning an invariant representation.
A related line of work also leverages inferred subgroup information to improve worst-case model perfo... | [
{
"start": 0,
"end": 717,
"label": 0
},
{
"start": 717,
"end": 837,
"label": 1
},
{
"start": 837,
"end": 1186,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | openreview | ICLR | 2,021 | clean_v2 | 9e0f95a67c44371e13fc3227ea21dc5e8bd4a30462646758a21887eb06a8fdda | human:9e0f95a67c44371e13fc3227ea21dc5e8bd4a30462646758a21887eb06a8fdda | or:9e0f95a67c44371e/p092 | null | 0 | 1,186 | 219 | 992 | null | null | stage2-epoch0-edit_claude-58 |
arxiv:33180d3fa9567581/p00/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 124 | Moreover, as a consequence of the mismatching boundary condition or equivalently of the different brane-tensions, the two branes intersect at an angle in the bulk, as illustrated in Fig. 7. Typically, in theories which allow for a quantum gravity UV-completion, there can be additional light degrees of freedom localized... | [
{
"start": 0,
"end": 352,
"label": 0
},
{
"start": 352,
"end": 476,
"label": 1
},
{
"start": 476,
"end": 1405,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | 33180d3fa9567581d70d51f4de541e01df03efc7406088498113178cc718321d | human:33180d3fa9567581d70d51f4de541e01df03efc7406088498113178cc718321d | arxiv:33180d3fa9567581/p00 | null | 0 | 1,405 | 0 | 906 | null | null | stage2-epoch0-edit_claude-59 |
a8edfa05bef6ea6a/p149/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 71 | In theory any type of environment feedback can inform the LLM planner, as long as it can be expressed through language. We focus on the specific forms of feedback shown in Fig 2: (1) task-specific feedback, such as success detection, and (2) scene-specific feedback (either “passive” or “active”), which describes the sc... | [
{
"start": 0,
"end": 497,
"label": 0
},
{
"start": 497,
"end": 568,
"label": 1
},
{
"start": 568,
"end": 1225,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | CoRL | 2,022 | paired_pdf_v1 | a8edfa05bef6ea6ad5021bf2bd63283c1eeb865452c81aa9f00566761c5f3f1e | paper:a8edfa05bef6ea6ad5021bf2bd63283c1eeb865452c81aa9f00566761c5f3f1e | null | null | 0 | 1,225 | 478 | 909 | null | null | stage2-epoch0-edit_claude-60 |
pes2o:dcb3da1d96654b67/p03/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 97 | PTEN mutation screening
The entire coding region of PTEN was screened. For each tumor and control samples, all nine exons were amplified. PCR products having altered mobility were directly sequenced. All sequence variants were confirmed by sequencing both strands from an independent PCR reaction. Based on our PTEN seq... | [
{
"start": 0,
"end": 201,
"label": 0
},
{
"start": 201,
"end": 298,
"label": 1
},
{
"start": 298,
"end": 488,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | dcb3da1d96654b67328e84af000c2f394f72ac1c4d852382a5c8a66de947df36 | human:dcb3da1d96654b67328e84af000c2f394f72ac1c4d852382a5c8a66de947df36 | pes2o:dcb3da1d96654b67/p03 | null | 0 | 488 | 25 | 488 | null | null | stage2-epoch0-edit_claude-61 |
pmc:a98ed35b26c18c3a/p02/opus-insert_one@614 | stage2-epoch0 | edit_claude | llm_edit | mixed | 97 | the parotid gland resulted in a few cases of PLN recurrence after treatment of NPC with IMRT [13](#cam41154-bib-0013){ref-type="ref"}, [14](#cam41154-bib-0014){ref-type="ref"}, [15](#cam41154-bib-0015){ref-type="ref"}, which suggested that potential or definite PLN metastasis in preliminarily diagnosed patients might b... | [
{
"start": 0,
"end": 879,
"label": 0
},
{
"start": 879,
"end": 976,
"label": 1
},
{
"start": 976,
"end": 1927,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | a98ed35b26c18c3a5da4301ac2c487d232c0dfca240e727b33b897abc4c97051 | human:a98ed35b26c18c3a5da4301ac2c487d232c0dfca240e727b33b897abc4c97051 | pmc:a98ed35b26c18c3a/p02 | null | 614 | 2,541 | 741 | 1,927 | null | null | stage2-epoch0-edit_claude-62 |
or:2024516c433ca9c3/p003/sonnet-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 91 | ABSTRACT
Multiple instance learning (MIL) is a machine learning paradigm which learns the mapping between bags of instances and bag labels. There are different MIL tasks which can be solved by different MIL methods. One common component of all MIL methods is the MIL pooling filter, which obtains bag level representati... | [
{
"start": 0,
"end": 785,
"label": 0
},
{
"start": 785,
"end": 876,
"label": 1
},
{
"start": 876,
"end": 1590,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | ICLR | 2,021 | clean_v2 | 2024516c433ca9c3baed5009bc8fb0ed83d763bf0cd7b9a223ceed8d6c0fe21f | human:2024516c433ca9c3baed5009bc8fb0ed83d763bf0cd7b9a223ceed8d6c0fe21f | or:2024516c433ca9c3/p003 | null | 0 | 1,590 | 10 | 1,587 | null | null | stage2-epoch0-edit_claude-63 |
icml-2017-0098c4f3ce0c0d90/q054/opus-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 117 | Proposition 18. For any f : D n → B with global sensitivity ∆ = sup D∼D 0 kf (D) − f (D 0 )k B , S ENSITIVI - ˆ ≤ ∆. As a result, TY S AMPLER ’s random sensitivity ∆ Algorithm 2 run with any of the sensitivity-induced private mechanisms of Corollaries 10–14 achieves utility dominating that of the respective mechanisms ... | [
{
"start": 0,
"end": 333,
"label": 0
},
{
"start": 333,
"end": 450,
"label": 1
},
{
"start": 450,
"end": 1214,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | proceedings | ICML | 2,017 | paired_pdf_v1 | icml-2017-0098c4f3ce0c0d90 | paper:icml-2017-0098c4f3ce0c0d90 | null | null | 0 | 1,214 | 333 | 1,100 | null | null | stage2-epoch0-edit_claude-64 |
pmc:de3e90ed3b0f93f1/p02/opus-rewrite_two | stage2-epoch0 | edit_claude | llm_edit | mixed | 195 | We thank the reviewers and editors for this suggestion. We are committed to help the community to use our method. In fact, we have already received and responded to many inquiries throughout the world after the manuscript was deposited in BioRxiv. All plasmids that have not been previously reported will be deposited to... | [
{
"start": 0,
"end": 1108,
"label": 0
},
{
"start": 1108,
"end": 1303,
"label": 1
},
{
"start": 1303,
"end": 1591,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | de3e90ed3b0f93f1820270b5c67310718497f20ed05a64658fcbb2daf3a7557f | human:de3e90ed3b0f93f1820270b5c67310718497f20ed05a64658fcbb2daf3a7557f | pmc:de3e90ed3b0f93f1/p02 | null | 0 | 1,591 | 741 | 1,591 | null | null | stage2-epoch0-edit_claude-65 |
pes2o:08f85b9149fbf3de/p01/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 96 | SNP Genotyping
SNP genotyping was performed on a set of non-redundant genotypes that differed to a lesser extent from that used in the SSR analysis due to technical reasons (highlighted in Table S1). The SNP set consisted of 47 SNP loci previously used and proposed for grapevine fingerprinting [38], to which the SNP l... | [
{
"start": 0,
"end": 532,
"label": 0
},
{
"start": 532,
"end": 628,
"label": 1
},
{
"start": 628,
"end": 1623,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pes2o | null | null | host_pes2o | 08f85b9149fbf3de0db677c9f5700070df21b43201e9b49225343d5ab35124f7 | human:08f85b9149fbf3de0db677c9f5700070df21b43201e9b49225343d5ab35124f7 | pes2o:08f85b9149fbf3de/p01 | null | 0 | 1,623 | 16 | 1,569 | null | null | stage2-epoch0-edit_claude-66 |
arxiv:b298a323f41767ad/p02/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 98 | These findings led us to the conclusion that an additional large-scale molecular gas tracer is needed besides CO, one that could provide an independent estimate of molecular gas content without being subject to the same set of properties and deficiencies as signals from the CO molecule. Desirable features of such a rel... | [
{
"start": 0,
"end": 1010,
"label": 0
},
{
"start": 1010,
"end": 1108,
"label": 1
},
{
"start": 1108,
"end": 1972,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | b298a323f41767ad6cc095ee9cadd8703de8d327ffbf4d697255690bfea9eacb | human:b298a323f41767ad6cc095ee9cadd8703de8d327ffbf4d697255690bfea9eacb | arxiv:b298a323f41767ad/p02 | null | 0 | 1,972 | 739 | 1,308 | null | null | stage2-epoch0-edit_claude-67 |
pmc:5429949405c4a6af/p01/opus-insert_one@1567 | stage2-epoch0 | edit_claude | llm_edit | mixed | 133 | a predominance of *N*. *americanus* worms may therefore have been another reason for the lack of an association between hookworm infection and anemia in this setting.
In all the models, statistically significant differences were seen between the clusters. This indicates that cluster-level factors, beyond the individua... | [
{
"start": 0,
"end": 257,
"label": 0
},
{
"start": 257,
"end": 390,
"label": 1
},
{
"start": 390,
"end": 1992,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 5429949405c4a6af4cf9d2f6c8909fb38d752a3b5377f6528af6ef0069bb5653 | human:5429949405c4a6af4cf9d2f6c8909fb38d752a3b5377f6528af6ef0069bb5653 | pmc:5429949405c4a6af/p01 | null | 1,567 | 3,559 | 168 | 857 | null | null | stage2-epoch0-edit_claude-68 |
pes2o:60c7d03d1251675b/p02/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 192 | e review of the measurement model was based on convergent validity, composite reliability, and the significance test. Convergent validity was measured by the loading factors that are the coefficients of correlation between the indicators and their latent constructs. Indicators that have high loading factors contribute ... | [
{
"start": 0,
"end": 1041,
"label": 0
},
{
"start": 1041,
"end": 1233,
"label": 1
},
{
"start": 1233,
"end": 1609,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | 60c7d03d1251675bad398579731863d2b09e58084880858530a225b74c44b1ef | human:60c7d03d1251675bad398579731863d2b09e58084880858530a225b74c44b1ef | pes2o:60c7d03d1251675b/p02 | null | 0 | 1,609 | 649 | 1,233 | null | null | stage2-epoch0-edit_claude-69 |
arxiv:7991f7c76d2b7ceb/p00/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 73 | Meta learning for optimization. Meta learning or learning to learn has a long history in psychology ward1937reminiscence ; lake2017building . Meta learning for optimization can date back to 1980s-1990s schmidhuber1992learning ; bengio1991learning , aiming to meta-learn the optimization process of learning itself. Recen... | [
{
"start": 0,
"end": 655,
"label": 0
},
{
"start": 655,
"end": 728,
"label": 1
},
{
"start": 728,
"end": 2185,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | 7991f7c76d2b7ceb2aecd3b986f13d8c90d7727a31212a7ad24131f86c3a0dd2 | human:7991f7c76d2b7ceb2aecd3b986f13d8c90d7727a31212a7ad24131f86c3a0dd2 | arxiv:7991f7c76d2b7ceb/p00 | null | 0 | 2,185 | 0 | 1,454 | null | null | stage2-epoch0-edit_claude-70 |
arxiv:7ee6293adc623aa5/p01/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 168 | One should pay attention to how the concept of yield stress is used in this work. For quasi-static loading simulations we use a typical way of defining it in DDD simulations, as the stress corresponding to a small amount of plastic strain. However, we also use ”true measures” of the yield stress which means that if the... | [
{
"start": 0,
"end": 1057,
"label": 0
},
{
"start": 1057,
"end": 1225,
"label": 1
},
{
"start": 1225,
"end": 1392,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | arxiv | null | null | host_arxiv | 7ee6293adc623aa55718152d0588903a2c7f017f1d76cb5a56eada947a4c2ed6 | human:7ee6293adc623aa55718152d0588903a2c7f017f1d76cb5a56eada947a4c2ed6 | arxiv:7ee6293adc623aa5/p01 | null | 0 | 1,392 | 669 | 1,392 | null | null | stage2-epoch0-edit_claude-71 |
pmc:41cef23049c2e4d8/p00/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 98 | The initiation of SBT after a period of MV is associated with some loss of lung aeration in critically ill subjects. Using the same LUS score technique as Bouhemad et al. This semi-quantitative score grades each lung region according to the ultrasound pattern observed. (lower scores = better aeration),^([@r25])^ Soumme... | [
{
"start": 0,
"end": 171,
"label": 0
},
{
"start": 171,
"end": 269,
"label": 1
},
{
"start": 269,
"end": 1522,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 41cef23049c2e4d824fa1a7526b4dd1dcdf4732c95d84ca8f13c37b7038e13f5 | human:41cef23049c2e4d824fa1a7526b4dd1dcdf4732c95d84ca8f13c37b7038e13f5 | pmc:41cef23049c2e4d8/p00 | null | 0 | 1,522 | 0 | 901 | null | null | stage2-epoch0-edit_claude-72 |
pmc:ba46aece9057bfac/p00/sonnet-rewrite_one@0 | stage2-epoch0 | edit_claude | llm_edit | mixed | 172 | In a recent pilot study, we tested the feasibility of imaging the DT-walking-related changes in brain glucose metabolism in patients with PD ([@B81]). Fifteen patients with PD (stage II and III) were scanned with FDG-PET. Ten patients were rested during the FDG uptake period and five patients performed DT treadmill wal... | [
{
"start": 0,
"end": 337,
"label": 0
},
{
"start": 337,
"end": 509,
"label": 1
},
{
"start": 509,
"end": 2403,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | ba46aece9057bfacf621205d0e171e7674f93a1db2964007e3433fd7872c30a9 | human:ba46aece9057bfacf621205d0e171e7674f93a1db2964007e3433fd7872c30a9 | pmc:ba46aece9057bfac/p00 | null | 0 | 2,403 | 0 | 1,273 | null | null | stage2-epoch0-edit_claude-73 |
f0edb68c6c96920e/p023/opus-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 168 | a masked autoencoder to learn node representations for heterogeneous graphs. Following VGAE, most of its variants only use node attributes in the encoder but do not decode node attributes. Though some other models (Mehta et al., 2019; Cheng et al., 2021) use separate decoder heads to generate graph edges and node attri... | [
{
"start": 0,
"end": 627,
"label": 0
},
{
"start": 627,
"end": 795,
"label": 1
},
{
"start": 795,
"end": 2035,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | openreview | TMLR | 2,022 | paired_pdf_v1 | f0edb68c6c96920e83765d058740b55e6d01451ca085b59f10e0ad9212b4b924 | paper:f0edb68c6c96920e83765d058740b55e6d01451ca085b59f10e0ad9212b4b924 | null | null | 0 | 2,035 | 463 | 1,134 | null | null | stage2-epoch0-edit_claude-74 |
pes2o:1e32596cd9e8f896/p04/opus-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 164 | In animal breeding, genetic differences in phenotypic stability can be detected as genetic heterogeneity of (within-family) residual variance, where a low variance indicates stable performance across environmental factors. Genetic heterogeneity of residual variance has been estimated for different traits in a number of... | [
{
"start": 0,
"end": 1873,
"label": 0
},
{
"start": 1873,
"end": 2037,
"label": 1
}
] | null | spans | anthropic/claude-opus-5-5 | pes2o | null | null | host_pes2o | 1e32596cd9e8f896516dc71f3d6412e920706a80e4fd8cb94e375ceee42dd8af | human:1e32596cd9e8f896516dc71f3d6412e920706a80e4fd8cb94e375ceee42dd8af | pes2o:1e32596cd9e8f896/p04 | null | 0 | 2,037 | 1,236 | 2,037 | null | null | stage2-epoch0-edit_claude-75 |
pmc:f5244f347d364cb8/p02/sonnet-rewrite_two@70 | stage2-epoch0 | edit_claude | llm_edit | mixed | 146 | between tree cover and DD downstream can vary across different socioeconomic contexts---as should be the case when there are multiple options for assuring health outcomes but those options vary across settings. On average, tree cover is statistically significant only in rural regions, where populations tend to be more ... | [
{
"start": 0,
"end": 1139,
"label": 0
},
{
"start": 1139,
"end": 1285,
"label": 1
},
{
"start": 1285,
"end": 2358,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | pmc | null | null | host_pmc | f5244f347d364cb8d5b5875adecb7dec3f020642176214e6a2d2821d1c3b2a74 | human:f5244f347d364cb8d5b5875adecb7dec3f020642176214e6a2d2821d1c3b2a74 | pmc:f5244f347d364cb8/p02 | null | 70 | 2,428 | 736 | 1,484 | null | null | stage2-epoch0-edit_claude-76 |
or:9497e215177a1807/p063/sonnet-insert_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 78 | m
We refer the reader to Theorem 16 in Appendix D for a formal statement and the proof. Notice that the above theorem implies that our proposed GNN + architecture for the above graph problems can indeed be trained using very few samples as opposed to the traditional GNN mp networks since the GNN + network requires muc... | [
{
"start": 0,
"end": 849,
"label": 0
},
{
"start": 849,
"end": 927,
"label": 1
},
{
"start": 927,
"end": 930,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | openreview | ICLR | 2,021 | clean_v2 | 9497e215177a1807691a859df8d6d2f257c89353a985669a9003e3e86e2585c0 | human:9497e215177a1807691a859df8d6d2f257c89353a985669a9003e3e86e2585c0 | or:9497e215177a1807/p063 | null | 0 | 930 | 3 | 927 | null | null | stage2-epoch0-edit_claude-77 |
arxiv:f159707ec32270a0/p04/sonnet-rewrite_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 85 | We tried to automate some of the tasks (e.g., filtering repositories that contain packages or finding the dependencies), using both PyPI and GitHub API, or the johnnydep tool. There were false positives and false negatives. This led us to manually cross check the results obtained from our automated search.
Whether a p... | [
{
"start": 0,
"end": 538,
"label": 0
},
{
"start": 538,
"end": 623,
"label": 1
},
{
"start": 623,
"end": 1229,
"label": 0
}
] | null | spans | anthropic/claude-sonnet-5-5 | arxiv | null | null | host_arxiv | f159707ec32270a0111212b87c8364e7e38ff0780a965fa1972785e43ae7a50c | human:f159707ec32270a0111212b87c8364e7e38ff0780a965fa1972785e43ae7a50c | arxiv:f159707ec32270a0/p04 | null | 0 | 1,229 | 309 | 696 | null | null | stage2-epoch0-edit_claude-78 |
pmc:93830116de3a2057/p00/opus-split_one | stage2-epoch0 | edit_claude | llm_edit | mixed | 139 | For instance, growth inhibition is seen in breast cancer cell lines following administration of caffeic acid phenethyl ester (CAPE) derived from propolis. CAPE modulates intra-tumor angiogenesis as well causes tumoral apoptosis thus decreasing tumor growth in breast carcinomas[@bib2]. It also attenuates tumor resistanc... | [
{
"start": 0,
"end": 527,
"label": 0
},
{
"start": 527,
"end": 666,
"label": 1
},
{
"start": 666,
"end": 1091,
"label": 0
}
] | null | spans | anthropic/claude-opus-5-5 | pmc | null | null | host_pmc | 93830116de3a2057a756a4be8d1ec1b4cda81cd7ab11542e5cee0c8085cd6f23 | human:93830116de3a2057a756a4be8d1ec1b4cda81cd7ab11542e5cee0c8085cd6f23 | pmc:93830116de3a2057/p00 | null | 0 | 1,091 | 0 | 666 | null | null | stage2-epoch0-edit_claude-79 |
End of preview. Expand in Data Studio
Open Text Detector training mix: Oct 7
The training mix behind woog/open-text-detector-qwen36-35b-a3b-oct10, the model that scores every paper on opentextdetector.com: a LoRA classifier on Qwen3.6-35B-A3B that labels each token, sentence and document of English text as human- or AI-written. The mix targets research papers, in particular small AI edits (one or a few sentences) inside otherwise human papers.
It holds 72,000 training rows, all distinct windows, seen once. They are stored as three files of 24,000 (split_file stage2-epoch0/1/2); the model read all three in one pass. The mix was built on Oct 7, 2026 from the same pools as T2.1, scaled to 72,000 windows with DeepSeek edits added and human paper text matched in format to the AI rows. The model trained on it skipped the earlier recipe's segment-only first stage, so there is no stage-1 split.
Splits
| Split | Rows | Use |
|---|---|---|
stage2 |
72,000 | All heads and LoRA, one pass. |
selection |
2,100 | Windows from held-out papers, used for the selection loss. Not trained on. |
calibration |
2,019 | Windows from further held-out papers, reserved for threshold calibration. Not trained on. |
The site's published cutoffs are set on 2,400 whole human papers instead; see the Calibration page.
Composition
Labels: human = no AI characters, AI = no human characters, mixed = both (an AI span inside human context).
| Key | Rows | Human / mixed / AI | Writers | Description |
|---|---|---|---|---|
edit_claude |
13,334 | 0 / 13,334 / 0 | claude-opus-5-5 6,623, claude-sonnet-5-5 5,802, claude-haiku-4-5 909 | One to a few sentences inserted or rewritten by Claude inside a human paper passage |
edit_luna |
8,132 | 0 / 8,132 / 0 | gpt-6-luna 8,132 | Sentence-level edits and splices by GPT-6 Luna inside human paper passages |
edit_deepseek |
3,356 | 0 / 3,356 / 0 | deepseek-v4.1-flash 3,356 | Sentence-level edits by DeepSeek V4.1 Flash inside human paper passages |
edit_gpt |
378 | 0 / 378 / 0 | gpt-6.1-sol 378 | Sentence-level edits by GPT-6.1 Sol inside human paper passages |
para_luna |
4,385 | 0 / 4,382 / 3 | gpt-6-luna 4,385 | A paragraph regenerated by GPT-6 Luna, in its original human context |
para_e |
3,505 | 0 / 3,280 / 225 | claude-sonnet-5-5 1,796, claude-opus-5-5 1,322, claude-haiku-4-5 217, gpt-6.1-sol 170 | A paragraph rewritten or drafted by Claude or GPT-6.1 Sol, in human context |
para_deepseek |
810 | 0 / 810 / 0 | deepseek-v4.1-flash 810 | A paragraph drafted by DeepSeek V4.1 Flash, in human context |
sections_ai |
4,200 | 0 / 978 / 3,222 | claude-opus-5-5 1,964, claude-sonnet-5-5 1,944, claude-haiku-4-5 292 | A paper section written by Claude from the paper's surrounding text |
mirrors |
4,200 | 0 / 0 / 4,200 | gpt-6-luna 4,200 | Fully AI documents written by GPT-6 Luna to mirror a human passage's genre and topic |
fullpapers |
1,500 | 0 / 0 / 1,500 | gpt-6.1-sol 775, gpt-6-luna 725 | Windows from full manuscripts written by GPT models from a human title and abstract (body only) |
claude_fp |
1,500 | 0 / 0 / 1,500 | claude-sonnet-5-5 640, claude-opus-5-5 588, claude-haiku-4-5 272 | Windows from full manuscripts written by Claude from a human title and abstract (body only) |
human_paper_orig |
9,000 | 9,000 / 0 / 0 | — | Unedited human paper text (paired originals and clean-v2 extractions) |
human_section |
3,000 | 3,000 / 0 / 0 | — | Human originals of the sections in sections_ai |
human_para_e |
3,000 | 3,000 / 0 / 0 | — | Human originals matched to para_e |
human_host |
3,000 | 3,000 / 0 / 0 | — | Unedited human host passages of the kind used for edits |
human |
5,100 | 5,100 / 0 / 0 | — | Generic human text from 34 non-paper and paper sources |
gradtex |
3,600 | 1,804 / 0 / 1,796 | — | GRADTEX (derived from MAGE); document-level labels only |
AI writers: openai/gpt-6-luna 17,442, anthropic/claude-opus-5-5 10,497, anthropic/claude-sonnet-5-5 10,182, deepseek/deepseek-v4.1-flash 4,166, anthropic/claude-haiku-4-5 1,690, openai/gpt-6.1-sol 1,323. GRADTEX AI rows come from the generators of its upstream sources and carry no per-row writer. No Claude Fable output is included; it is reserved for evaluation.
Small edits. In mixed rows, the AI span has a median of 210 characters (5th–95th percentile 76–1,184); 16,590 rows have under 200 AI characters.
Sources of the human text
- Research papers, all dated 2022 or earlier: NeurIPS, ICML and ACL proceedings, 2013–2021 (
source=proceedings); OpenReview papers from ICLR, NeurIPS, CoRL and TMLR (openreview); and open-access passages from PMC, arXiv, peS2o and the ACL Anthology (pmc,arxiv,pes2o,acl).venueandyearare filled where known. - Generic human text (
human, and the seeds ofmirrors): gutenberg 971, wikipedia 899, writingprompts 713, stack_nontech 666, pmc 597, amazon2018 566, cccc 561, arxiv 559, pes2o 377, acl 296, stack_tech 252, voa 231, foodista 204, scp 198, ubuntu_irc 196, globalvoices 182, wikibooks 173, wikivoyage 166, pressbooks 156, oanc_slate 154, wikisource 134, govreport 123, libretexts 115, wikinews 115, imdb 112, opinrank 107, wiki_talk 104, scidev 101, worldbank 91, fed 87, oanc_icic 50, eff 44. - Full-manuscript seeds: the human title and abstract of an existing paper. Only the generated body is included, so no seed text appears.
fullpapersseeds are ICLR 2026, NeurIPS 2025, ICML 2025 and COLM 2025 papers;claude_fpseeds are half from 2022 or earlier and half from those 2025–2026 venues.venueandyeargive the seed paper. - GRADTEX, as published upstream.
Construction
- Long documents were cut into non-overlapping windows that fit 510 tokens under the Qwen3.5, Qwen3.6 and Gemma tokenizers. For paired and edited rows the window is placed around the AI span.
- Stage-2 rows were screened against every held-out evaluation list: rows from evaluation, selection or calibration papers, or sharing a sentence of 10+ words with them, were dropped.
- Allocation per key was round-robin over papers. When a key ran short it was filled from another key in the same group (for example
edit_gpthad only 378 usable rows).
Fields
| Field | Meaning |
|---|---|
id |
Row id; encodes the source document, edit and window offset. Unique in each split. |
split_file, draw_id |
Original training file and draw position. |
dataset, kind |
Mix key (table above) and row type. |
label |
human, mixed or ai, derived from regions (or document_label). |
text |
The window. |
regions |
Character spans of text with label 1 = AI, 0 = human. The source of the token, sentence and segment labels. |
document_label, supervision |
GRADTEX rows only: document_only supervision with a document label (1 = AI). Their regions are empty. |
ai_chars |
Number of AI-labelled characters. |
generator |
Writer of the AI text; null for human-only rows and GRADTEX. |
source, venue, year |
Origin of the human text or seed (see above). |
text_format |
Text extraction lineage: paired_pdf_v1, clean_v2, archive_v1 (paper PDF extractions), host_* (open-access host passages), generic_human, generated_doc, generated_markdown, gradtex. |
paper_id, group, host_id, seed_id |
Source identifiers. Split by group to keep a document's windows together. |
source_start, source_end, target_start, target_end |
Window offsets in the source document, and the span around the edit within the window. |
original_split, protected |
Bookkeeping from the upstream paired-paper split. |
Known limitations
- Text format correlates with the label. In stage 2, the share of non-human rows is 0.51 for
clean_v2(16,750 rows); 0.77 forpaired_pdf_v1(23,037 rows); 0.69 forhost_pmc(7,790 rows); 0.78 forhost_arxiv(3,818 rows); 0.72 forhost_pes2o(4,226 rows). A model can partly learn format as a shortcut. - GRADTEX labels are document-level only.
- Some Claude Opus 5.5 rows may be Claude Opus 5. When a safety classifier stopped a writing agent, it continued on Claude Opus 5 while its output kept the original label. 124 rows (117
para_e, 7edit_claude) come from batches where this happened; they are AI-written either way. - The Claude Haiku rows were written by Claude Haiku 4.5, not Haiku 5.5 as intended.
stage2contains 218 rows whose text duplicates another row.
Rights
This collection combines material under different terms and grants no new rights to any part of it.
- Human text keeps the terms of its source: per-paper terms for conference papers, per-article open-access licenses for PMC, arXiv, peS2o and ACL Anthology text, CC BY-SA for Wikimedia projects, and the release terms of each research corpus (including GRADTEX).
- AI-written text is subject to the terms of the providers of the listed models.
Please open a discussion to request removal of any item.
- Downloads last month
- 74
Models trained or fine-tuned on woog/open-text-detector-mix-oct7