AI Paper Atlas

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

FormalML: A Benchmark for Evaluating Formal Subgoal Completion in Machine Learning Theory

Xiao-Wen Yang, Zihao Zhang, Jianuo Cao, Zhi Zhou, Zenan Li, Lan-Zhe Guo, Yuan Yao, Taolue Chen, Yu-Feng Li, Xiaoxing Ma

ICLR 2026 · Large language models › Agents, code and math · approach: Retrieval-augmented formal theorem proving · 3 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
add 10 candidate premises to the prompt (ablation)DeepSeek-Prover-V2 (noCoT) · FormalML retrieval subsetPass@32 (%): 58.37 -> 69.29Improved
add 10 candidate premises to the prompt (ablation)DeepSeek-Prover-V2 (CoT) · FormalML retrieval subsetPass@32 (%): 37.94 -> 47.71Improved
add 10 candidate premises to the prompt (ablation)STP · FormalML retrieval subsetPass@32 (%): 57.72 -> 56.43Hurt
add 10 candidate premises to the prompt (ablation)Goedel-Prover · FormalML retrieval subsetPass@32 (%): 39.04 -> 41.05Improved
add 10 candidate premises to the prompt (ablation)DeepSeek-Prover-V1.5 · FormalML retrieval subsetPass@32 (%): 49.45 -> 46.86 (Pass@16); 52.94 -> 53.59 (Pass@32)Mixed
increase candidate premises from 10 to 20 (ablation)provers in general · FormalML retrieval subsetPass@32 (%): DeepSeek-Prover-V2 (noCoT) 69.29 -> 65.80Hurt
use long-CoT reasoning (DeepSeek-Prover-V2 CoT vs noCoT) (ablation)DeepSeek-Prover-V2 · FormalML (All)Pass@32 (%): 62.06 -> 39.50Hurt
increase sample budget from 1 to 32 (ablation)Kimina-Prover-Preview-7B · FormalML (All)Pass@K (%): 15.43 -> 45.74Improved
best-first tree search with larger budget (8x50 vs 4x20) (ablation)BFS-Prover · FormalML (All)Pass@K (%): 22.26 -> 25.31No clear effect
one round of expert iteration on mined subgoal problemsDeepSeek-Prover-V2 (noCoT) and Goedel-Prover-V2-8B · FormalML (All)Pass@1: Pass@1 improved by about 20%Improved
increase proof difficulty level (proof length 1 to 5) (ablation)all whole-proof provers · FormalMLPass@KHurt

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