AWhat changed
V1 iteratively synthesized formal statements/proofs with filtering and parallel negation checks; V1.5 added proof-assistant-feedback RL and intrinsic-reward MonteCarlo search with truncate-and-resume proof completion.
Scientific scope
V1 iteratively synthesized formal statements/proofs with filtering and parallel negation checks; V1.5 added proof-assistant-feedback RL and intrinsic-reward MonteCarlo search with truncate-and-resume proof completion. One two-source method family distinct from already formal ProverV2 recursive decomposition. Inherits DeepSeekMath, Lean and expert iteration. Source-specific initials retained; autoformalization fidelity is not implied by valid proofs. Sampling and search budgets are part of results.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. One two-source method family distinct from already formal ProverV2 recursive decomposition. Inherits DeepSeekMath, Lean and expert iteration. Source-specific initials retained; autoformalization fidelity is not implied by valid proofs. Sampling and search budgets are part of results.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- DeepSeek-Prover: Advancing Theorem Proving in LLMs through Large-Scale Synthetic Data
- DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction