RLHF is just barely RL

Re:RLHF is just barely RL by Barrin92 on Thursday August 08, 2024 @08:16AM (#41190723)

> but I'm surprised to hear someone think that it's not even "barely usable".

write performance oriented and memory safe C++ code. Current coding assistants are glorified autocomplete for unit tests or short api endpoints or what have you but if you have to write any safety oriented code or you have to think about what the hardware does it's unusable.

I tried using several of the assistants and they write broken or non-performant code so regularly it's irresponsible to use them.

[ Parent | on HN ]

Re:RLHF is just barely RL by agos on Thursday August 08, 2024 @08:21AM (#41190767)

I've also had trouble having assistants help with CSS, which is ostensibly easier than performance oriented and memory safe C++

[ Parent | on HN ]

Re:RLHF is just barely RL by imtringued on Thursday August 08, 2024 @10:23AM (#41191855)

Isn't this a good reward function for RL? Take a codebase's test suite. Rip out a function, let the LLM rewrite the function, benchmark it and then RL it using the benchmark results.

[ Parent | on HN ]