3 ms·
I suppose the issue is a data problem, there being relatively little high quality data explaining how things should be solved in binary. As such making the lear
by koe123 2y ago
I suppose the issue is a data problem, there being relatively little high quality data explaining how things should be solved in binary. As such making the learning mapping between prompt (english) and good solution (binary) difficult.
- ninetyninenine 2y agoThe data is easily generated by compiling the code.
- 0xCMP 2y agoBut compiled code loses a lot of the "extra" data. Also these are "language" models so I would be surprised if training on binaries was much more efficient versus writing in some kind of language. Besides, how do you even check the result now without running untrusted code? Every run of the model you need to reverse-engineer the binary?
- asdff 2y agoYou don’t even need a training set imo just enough permutations to get a distribution of putative functions and outputs.