Most AI benchmarks test retrieval — can a model find the known answer?
Most AI benchmarks test retrieval — can a model find the known answer? However, the hardest problems in science require discovery, can a system earn an answer nobody has yet? Meet TRACES — the world's first benchmark for measuring discoverative AI: AI that can
In scientific discovery, the answer is often not hidden, it simply does not exist yet.
In scientific discovery, the answer is often not hidden, it simply does not exist yet. So 𝗧𝗥𝗔𝗖𝗘𝗦 doesn't grade against an answer key. It scores the investigation. Today, the first three letters, each with one experiment behind it: T — one harness's entir