ojhunt wrote: > > Please review the llvm AI policy, and confirm you are following it. > > I did read the AI Tool Use Policy, and my understanding was that this > workflow was consistent with it. If I’ve misunderstood any part of it, I’d > appreciate clarification. > > To clarify my workflow: > > * I use LLMs to help find issues that I find interesting and that don’t > appear to have an existing fix. > * I manually reproduce the reported issue locally to check whether it still > exists or has already been fixed. > * I use LLMs to help narrow down where the problem might originate. > * I then write the code and run tests, and use LLMs for an additional local > code review. > * When writing PR descriptions, I use LLMs to translate my writing into > English. > > Could you clarify which part of this workflow raises concerns under the > policy, and what I should change? Does this level of assistance require an > Assisted-by tag or another form of disclosure? If so, I’ll make sure to > include it going forward. Thanks for the guidance!
Ok, my concern is that you have filed a large number of PRs against a wide array of different areas in llvm (libc and clang). Using an LLM for translation is 100% ok. I've made the "This PR doesn't appear to comply with our policy on tool-generated content" on all of your PRs simply due to the number, but if you have not used an LLM for those PRs then just reply referencing your reply here and it's all ok. It would be better to focus on a single area to build up experience and knowledge rather than trying to learn everything in the code base all at once. https://github.com/llvm/llvm-project/pull/229612 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
