Sample report
What a full paper check gives you
This is a real report, produced by uploading a short demonstration paper. We wrote the paper ourselves and put five kinds of mistake in it on purpose, so you can see how each one is reported.
An invented article in an invented journal, of the kind AI writing tools produce.
A classic 1974 paper listed as 1984.
Cited as if it were still valid evidence.
One sentence says far more than the cited paper does.
One paper is cited but not listed, another is listed but never cited.
demonstration-paper.docx
10 references · citation style detected: APA
Reference check
Each reference was looked up in CrossRef and OpenAlex, including their retraction records (from Retraction Watch). This confirms the work exists, its details match and it hasn't been retracted.
Corrected references are written from the published record, not from your text. Read each one before you use it, and add italics where your style asks for them. 2 references are left out because no usable record was found or the work was retracted.
Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children
Wakefield, A. J., Murch, S. H., Anthony, A., et al. (1998)
- This work has been retracted in 2010 (notice: https://doi.org/10.1016/s0140-6736(10)60175-4). Retracted work should not be cited as evidence.
Emergent humanlike reasoning in very large transformer language models: a longitudinal synthesis
Hartwell, P. R. (2021)
- No published work with this title was found in CrossRef or OpenAlex
Judgment under uncertainty: Heuristics and biases
Tversky, A., Kahneman, D. (1984)
- Year is 1984 in your paper but 1974 in the published record
Tversky, A., & Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/science.185.4157.1124
BERT: Pre-training of deep bidirectional transformers for language understanding
Devlin, J., Chang, M.-W., Lee, K., Toutanova, K. (2019)
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), 4171–4186. https://doi.org/10.18653/v1/n19-1423
Why most published research findings are false
Ioannidis, J. P. A. (2005)
Ioannidis, J. P. A. (2005). Why Most Published Research Findings Are False. PLoS Medicine, 2(8), e124. https://doi.org/10.1371/journal.pmed.0020124
Prospect theory: An analysis of decision under risk
Kahneman, D., Tversky, A. (1979)
Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), 263. https://doi.org/10.2307/1914185
Deep learning
LeCun, Y., Bengio, Y., Hinton, G. (2015)
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
Mastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., et al. (2016)
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., Dieleman, S., Grewe, D., Nham, J., Kalchbrenner, N., Sutskever, I., Lillicrap, T., Leach, M., Kavukcuoglu, K., Graepel, T., & Hassabis, D. (2016). Mastering the game of Go with deep neural networks and tree search. Nature, 529(7587), 484–489. https://doi.org/10.1038/nature16961
Attention is all you need
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., Polosukhin, I. (2017)
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention Is All You Need. arXiv (Cornell University), 30, 5998–6008. https://doi.org/10.48550/arxiv.1706.03762
Fabrication and errors in the bibliographic citations generated by ChatGPT
Walters, W. H., Wilder, E. I. (2023)
Walters, W. H., & Wilder, E. I. (2023). Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific Reports, 13(1). https://doi.org/10.1038/s41598-023-41032-5
Claim check: does the cited paper support your sentence?
For 5 of the 8 references we could find cited in your text, we compared the sentences that cite them (6 in all) with the cited paper's abstract (not its full text). Treat each verdict as a pointer: a claim can be supported in the body of a paper without appearing in its abstract, so open the paper before you change anything. 1 could not be checked because no abstract is available.
Attention Is All You Need
Not supported by the abstractYour sentence: It has been proved that transformers reason exactly like human experts in every domain (Vaswani et al., 2017).
Why: Abstract makes no claim about human-like reasoning across all domains.
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Can't tell from the abstractYour sentence: Building on this, bidirectional pre-training showed that a single model could be fine-tuned for many language tasks (Devlin et al., 2019).
Why: Abstract provided only contains citation metadata, no content about fine-tuning claim.
Why Most Published Research Findings Are False
Supported by the abstractYour sentence: The scale of irreproducible work is hard to measure, although one widely cited estimate suggests that most published research findings are false (Ioannidis, 2005), and large replication efforts have since examined this directly (Nosek, 2016).
Why: Abstract directly argues most published research findings are false.
From the abstract: “most current published research findings are false”
Prospect Theory: An Analysis of Decision under Risk
Supported by the abstractYour sentence: Their later work on prospect theory described how people evaluate gains and losses relative to a reference point rather than in absolute terms (Kahneman & Tversky, 1979).
Why: Abstract describes value assigned to gains and losses relative to a reference rather than absolute.
From the abstract: “value is assigned to gains and losses rather than to final assets”
Attention Is All You Need
Supported by the abstractYour sentence: The transformer architecture then removed recurrence entirely and relied only on attention mechanisms, which made training far more parallelisable (Vaswani et al., 2017).
Why: Abstract confirms transformer removes recurrence and is more parallelizable.
Fabrication and errors in the bibliographic citations generated by ChatGPT
Supported by the abstractYour sentence: One longitudinal study claimed that very large models develop humanlike reasoning as they grow (Hartwell, 2021), and fabricated references have been reported repeatedly in generated text (Walters & Wilder, 2023).
Why: Abstract reports fabricated citations occurring repeatedly in ChatGPT-generated text.
From the abstract: “fabricated bibliographic citations that do not represent actual scholarly works”
Citations in the text vs the reference list
We found 11 author-year citations in the text and compared them with the list. These are things to check by eye: the matching is by first author and year (or number), so an unusual format can cause a false alarm.
Listed but not cited in the text (2)
Remove these from the list, or cite them where they are used.
- Silver (2016): Mastering the game of Go with deep neural networks and tree search
- Tversky (1984): Judgment under uncertainty: Heuristics and biases
Cited in the text but not in the list (2)
Add the missing entry, or correct the citation.
- Tversky, 1974 (this author is in the list under a different year)
- Nosek, 2016
Run this on your own paper
One full reference check is free every month, with no card. The claim check shown above comes with paid paper checks and Professional.
Check my paper free
