Home

Awesome

ArxivLeaks

Most of papers on arxiv have latex files, which often contain much comment out. Dig up the valuable comments!

For example, you can extract a secret comment from "Attention Is All You Need", as below:

\paragraph{Symbol Dropout} In the source and target embedding layers, we replace a random subset of the token ids with a sentinel id. For the base model, we use a rate of $symbol\_dropout\_rate=0.1$. Note that this applies only to the auto-regressive use of the target ids - not their use in the cross-entropy loss.

We found "Symbol Dropout", which do not appear in the paper (pdf).

Run

Feed a text/page containing arxiv URLs by -t.

python -u run.py -t deepmind.html -s arxiv_dir

To test, run sh test.sh. This pre-downloads a publication page of deepmind.

You can also read only selected papers by -i, feeding their arxiv ids.

python -u run.py -i 1709.04905 1706.03762 -s arxiv_dir

Requirement

For Writers

You can remove %-comments from your file as follows:

perl -pe 's/(^|[^\\])%.*/\1%/' < old.tex > new.tex

This one line command is given from arxiv.