If I understand the results tables on repo correctly, their largest model achieves ~68% re-executability rate on code compiled with the q0 optimization flag. I'm unsure if that just tests if the decompiled code can be recompiled and executed, or if the programs need to produce the same result on some test cases. If the model is used to refine Ghidra outputs (I'm guessing this is some well-known decompilation framework) it can be used to achieve ~80% re-executability rate, which is better than Ghidra's baseline of ~34%.
- Posts
- 3
- Comments
- 302
- Joined
- 1 yr. ago
- Posts
- 3
- Comments
- 302
- Joined
- 1 yr. ago
- JumpDeleted
Permanently Deleted
- JumpDeleted
Permanently Deleted
- JumpDeleted
Permanently Deleted
- JumpDeleted
Permanently Deleted
- JumpDeleted
Permanently Deleted
Many distros have pretty comprehensive and up to date wikis (Arch and Debian for example). LLMs have a lot of out of date info as well, even the big proprietary ones. I can see how they might help a newcomer though, with how bad search engines are now (ironically, a lot of the problems are from AI generated articles SEO'd to hell).