Is Q* about solving maths problems?
➕
Plus
25
Ṁ3100
2027
50%
chance

Is the goal of Q* to solve maths problems?

If an Open-AI-produced document indicates that Q* is a system that primarily aims to solve maths problems this question will resolve to yes. If no such communication is given or the project is never released (by 30/1/2027 11:59pm) or it transpires that this project does not exist, this question will resolve to No.

"Primarily"

If Q* is a general system which among other things is good at solving maths problems (e.g. a better LLM), the question will resolve to No.

If the project's primary benchmark focuses on mathematical problem-solving, the question will resolve to Yes.

If any ambiguity remains, I'll leave the market open until no further clarification is hopefully needed. Although it can be hard to give precise rules about what it means to be "focused on maths problem-solving", I assume that it will be self-evident in practice once/if we know more about the system.

Get Ṁ1,000 play money
Sort by:

"improved stem reasoning" is quite close to math problems, and a big part of the benchmarks were about math, and a big part of the RL training was about math problems and mathematical lines of reasoning. not sure that counts as YES.

also openai will never admit o1 is downstream of Q*, but I really think that shouldn't make this market resolve NO. turning all "will X happen" type questions into "will company officially declare X happened" makes questions resolve NO way more often than would prescribe the spirit of the market imo.