How openai/primegaps186 works
This repository combines a conditional Lean proof of a prime-gap bound with a standalone Python numerical certificate. The main workflow is: assume three analytic/numerical inputs, derive DHL[40,2], apply it to the documented diameter-186 tuple, and conclude the liminf bound. Separately, the Python script recomputes a numerical trial and writes a receipt; the README cautions that this does not discharge the Lean axioms. Lean source details and certificate internals are only partially sampled, so the graph avoids unsupported internal call relationships.
Main entry points: prime_gap_186_certificate.py Numerical recomputation · PrimeGaps186.lean DHL theorem
Components 8
Lean formalization
- DHL theorem
PrimeGaps186.lean - Explicit tuple
PrimeGaps186.lean - Gap bound
PrimeGaps186.lean
Numerical certificate
- Numerical recomputation
prime_gap_186_certificate.py - JSON receipt
Assumed inputs
- Input axioms
PrimeGaps186.lean - Published estimates
Other components
- Project user
Connections 8
- Project user → Numerical recomputation: runs script
README.md - Numerical recomputation → JSON receipt: writes receipt
README.md - Project user → DHL theorem: checks proof
README.md - Published estimates → Input axioms: motivates estimates
README.md - Input axioms → DHL theorem: assumed inputs
README.md - DHL theorem → Explicit tuple: applies theorem
README.md - Explicit tuple → Gap bound: yields bound
README.md - Numerical recomputation → Input axioms: checks numerical input