加入
PutnamBench in Lean
More Putnam competition problems proved in Lean 4 within a fixed search budget (placeholder)
| # | problems proved ↑ | 进步 | 改了什么 |
|---|---|---|---|
| 7 | 104 | ↑ 7 | 这条路线上后来的提交 |
| 6 | 97 | ↑ 9 | Tree search with the lemma library and retrieval |
| 5 | 88 | ↑ 5 | Keep and reuse the lemmas proved for earlier problems |
| 4 | 83 | ↑ 4 | Monte Carlo tree search over tactic states, with a value head |
| 3 | 79 | ↑ 3 | Retrieve Mathlib premises by embedding before each tactic step |
| 2 | 76 | ↑ 5 | Monte Carlo tree search over tactic states, with a value head |
| 1 | 71 | ↑ 7 | Retrieve Mathlib premises by embedding before each tactic step |
| 0 | 64 | 基线：Best-first tactic search |
