BWhat changed
Released sparse Lean4-specialized agent models with language-server tooling and FLTEval repository tasks; the1.5stage describes CISPO reinforcement learning over multiturn proof/disproof and filesystem code-agent environments.
Scientific scope
Released sparse Lean4-specialized agent models with language-server tooling and FLTEval repository tasks; the1.5stage describes CISPO reinforcement learning over multiturn proof/disproof and filesystem code-agent environments. Source-reported model/workflow/resource, not invention of Lean, CISPO or proof engineering. Statement preservation, permitted axioms, compiler/kernel checks and natural-language intent are separate. Costs/pass budgets are configuration-dependent; benchmarks and bug reports were not replayed. Launch-era API availability is not asserted current.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Source-reported model/workflow/resource, not invention of Lean, CISPO or proof engineering. Statement preservation, permitted axioms, compiler/kernel checks and natural-language intent are separate. Costs/pass budgets are configuration-dependent; benchmarks and bug reports were not replayed. Launch-era API availability is not asserted current.
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.
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