The Automated LLM Speedrunning Benchmark measures how well LLM agents can reproduce previous innovations and discover new ones in language modeling. - View it on GitHub
Star
97
Rank
284920