logoalt Hacker News

in-silicoyesterday at 11:22 PM1 replyview on HN

Did the model really need to hack huggingface to get access to ExploitGym data? I'd imagine that once it had full internet access it could have just used the HF API or website (but the heavy prompting/nudging towards hacking made it do things the hard way).


Replies

Destructotortoday at 2:49 PM

I would be surprised if the solutions for the benchmark were publically available like that?