Fix fabricated citations + add Transformer ref & RTD API links - #22
Merged
Conversation
The Transformer/StateNet/AdapTrans docstrings carried a hallucinated author
list (Rançon, Bornschein, King, Schnupp, Willmore). Corrected to the real
authorship (verified online):
- Rançon, Masquelier & Cottereau (2025), Communications Biology 8:1456,
doi:10.1038/s42003-025-08858-3 (Transformer + StateNet)
- Rançon, Masquelier & Cottereau (2024), PLOS Comput. Biol. 20(8):e1012288,
doi:10.1371/journal.pcbi.1012288 (AdapTrans; also fixed the title to the
published one)
README_models.md:
- added the missing Transformer reference (Vaswani et al. 2017 architecture;
designed as the attention baseline in Rançon et al. 2025)
- linked every model's class to its RTD API page
- fixed two wrong years (DNet: Rahman 2016->2019; ConvNet2D: Pennington &
David 2023), both verified online
- de-linked two broken empty () links in the LN section
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Hsu et al. (2004) title (crcns_aa1): 'Modulation and phase spectrum … enhance neural discrimination' -> the published title 'Modulation power and phase spectrum … enhance neural encoding' (J. Neurosci. 24(41):9201-9211). - ns1: the DNet companion paper (doi pcbi.1006618) was labelled 'Rahman et al. 2018 PLoS Comp Biol'; the PLOS CB paper is 2019 (the 2018 was the bioRxiv preprint). Now consistent with the rest of the file. Full citation audit done this session: ~25 references verified online; all others (Machens 2004, Asari & Zador 2009, Lopez-Espejo 2019, Hsu/Borst 2004, Woolley 2005, Gill 2006, Amin 2010, Singh & Theunissen 2003, Elie & Theunissen 2015/2019, Rahman 2020 PNAS, Harper 2016, Brennan 2019, Brodbeck 2023, Bhattasali 2020, Pennington 2022/2023, Glasberg & Moore 1990, SincNet 2018, LEAF 2021, Wingert 2026, Drakopoulos 2025, Le 2025) confirmed correct. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Caught by Ulysse: hallucinated author lists in the model docstrings. All
citations below were verified online before writing.
Fabricated author list — fixed
Transformer(audio_zoo.py),StateNet(audio_zoo.py), andAdapTrans(
prefiltering.py) docstrings cited "Rançon, Bornschein, King, Schnupp,Willmore" — fabricated. Correct authorship:
doi:10.1038/s42003-025-08858-3 — Transformer + StateNet.
20(8):e1012288, doi:10.1371/journal.pcbi.1012288 — AdapTrans (also fixed
the title from the bioRxiv "theoretical framework" wording to the published
"A general model unifying …").
README_models.md
architecture; designed as the attention baseline in Rançon et al. (2025).
Rahman 2016 → 2019(PLOS CB 15(5):e1006618); ConvNet2D →
Pennington & David (2023)(PLOS CB 19(5):e1011110).
()links in the LN section.No code logic changed (docstrings + docs only).
🤖 Generated with Claude Code