AI Paper Atlas

Research map / Large language models / Agents, code and math

Propose, Solve, Verify: Self-Play Through Formal Verification

Alex Wilf, Pranjal Aggarwal, Bryan Parno, Daniel Fried, Louis-Philippe Morency, Paul Pu Liang, Sean Welleck

ICML 2026 · Large language models › Agents, code and math · approach: Code generation benchmarks · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train only on formally verified solutions (add verification filter) (ablation)PSV-Verus · Dafny2Verus / MBPP / HumanEval (transfer)pass@1: 31.82 -> 65.63 on Dafny2VerusImproved
add difficulty labels to proposer prompt (ablation)PSV-Verus · all three benchmarksaverage pass@k: 40.87 -> 43.31Improved
refresh proposer in-context examples each iteration (diversity) (ablation)PSV-Verus · all three benchmarksaverage pass@k: 38.71 -> 43.31Improved
filter proposed specs with a spec verifier (ablation)PSV-Verus · spec filteringinference compute: 2.1x increase in inference compute without itNo clear effect
scale questions per iteration from 4k to 32k (ablation)PSV-Verus · Dafny2Verus (TTT)pass@1: 59.5% -> 74.3%Improved
scale questions per iteration from 1k to 32k (ablation)PSV-Verus · MBPP (TTT)pass@1: 22.3% -> 44.3%Improved
scale questions per iteration from 1k to 32k (ablation)PSV-Verus · HumanEval (TTT)pass@1: 13.9% -> 26.3%Improved
split a 1k question budget over 5 iterations instead of 1 (ablation)PSV-Verus · Dafny2Verus (TTT)pass@1: 21.0% -> 42.3%Improved
split a 4k question budget over more iterations (ablation)PSV-Verus · Dafny2Verus (TTT)pass@1: 28.0% -> 46.9%Improved
replace baselines with PSV self-play proposing new specsRFT (expert iteration) · MBPP (TTT)pass@1: 3.83 -> 36.78Improved
train on Dafny2Verus with PSV then test on held-out MBPPPSV-Verus vs AlphaVerus · MBPP (transfer)pass@1: 6.48 -> 25.25Improved
difficulty prompting for proposer targetsPSV proposer · proposed question pass ratesmean pass rate: easy 0.55, medium 0.45, hard 0.36Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy